Nvidia Gives Free Cloud GPUs to Colleges to Accelerate AI Credentials
NVIDIA
Nvidia absorbs academic infrastructure costs through cloud-hosted GPUs and turnkey teaching kits, allowing universities to graduate certified artificial intelligence practitioners without purchasing local server labs.
NVIDIA's Deep Learning Institute (DLI) provides cloud GPU-accelerated coursework on an Open edX-based platform, enabling university faculty to deliver intensive workshops without managing local hardware infrastructure. (0:04)
The DLI Ambassador Program certifies academic faculty and institutional HPC staff to teach workshops for free, while excluding students and PhD candidates due to high turnover rates. (7:25)
NVIDIA distributes seven modular, full-semester Teaching Kits under non-commercial Creative Commons licenses, co-developed with top academic institutions across fields like deep learning, robotics, and generative AI. (10:12)
ECPI University demonstrated practical adoption by running 8-hour co-curricular Friday workshops with a roughly 70% student pass rate, integrating AI credentials across technical and non-technical disciplines. (32:55)
Detailed notes
## Background
Published on YouTube by NVIDIA on April 7, 2026, this presentation details NVIDIA's developer training initiatives designed to empower higher education faculty and researchers. According to the video description, the session covers downloadable teaching kits, a cloud-based GPU platform, and technical certification exams across AI and HPC domains, spotlighting university integrations such as ECPI's hands-on student projects.
## TL;DR
The presentation outlines how NVIDIA's Deep Learning Institute (DLI) provides higher education educators with turnkey, cloud-accelerated tools through the high-touch DLI Ambassador Program and self-serve Teaching Kits. By supplying browser-based cloud GPU access, auto-graded labs, and modular open-license courseware, the program eliminates local hardware barriers for institutions teaching deep learning, data science, and generative AI. ECPI University professor Paul Nussbaum illustrates this model in practice, detailing how intense, co-curricular eight-hour workshops yield verifiable credentials and engage diverse student cohorts. The talk concludes with an interactive Q&A addressing modular scheduling, staff eligibility criteria, and international cross-disciplinary degree integration.
## Atmosphere & style
The session carries an informative, collegial, and pragmatic conference tone, blending structured corporate program overviews with candid academic field reports. Paul Nussbaum injects energetic, relatable humor into the discussion, describing the thrill students experience when trading underpowered laptops for cloud GPUs that run like 'driving a Ferrari.' The concluding Q&A feels vibrant and collaborative, driven by genuine audience curiosity from global educators and innovators eager to overcome resource constraints and democratize AI training.
## Bulletin
- [00:04](https://www.youtube.com/watch?v=DIIteghPGx4&t=4s) NVIDIA's Deep Learning Institute (DLI) provides cloud GPU-accelerated coursework on an Open edX-based platform, enabling university faculty to deliver intensive workshops without managing local hardware infrastructure.
- [07:25](https://www.youtube.com/watch?v=DIIteghPGx4&t=445s) The DLI Ambassador Program certifies academic faculty and institutional HPC staff to teach workshops for free, while excluding students and PhD candidates due to high turnover rates.
- [10:12](https://www.youtube.com/watch?v=DIIteghPGx4&t=612s) NVIDIA distributes seven modular, full-semester Teaching Kits under non-commercial Creative Commons licenses, co-developed with top academic institutions across fields like deep learning, robotics, and generative AI.
- [32:55](https://www.youtube.com/watch?v=DIIteghPGx4&t=1975s) ECPI University demonstrated practical adoption by running 8-hour co-curricular Friday workshops with a roughly 70% student pass rate, integrating AI credentials across technical and non-technical disciplines.
## Key takeaways
- [00:04](https://www.youtube.com/watch?v=DIIteghPGx4&t=4s) DLI courses run on a customized Open edX platform wired directly to cloud GPUs, eliminating local hardware requirements for browser-based lab experiments.
- [00:04](https://www.youtube.com/watch?v=DIIteghPGx4&t=4s) Automated grading within DLI labs verifies student competency and issues verifiable, named certificates for resumes and professional profiles.
- [01:54](https://www.youtube.com/watch?v=DIIteghPGx4&t=114s) The Fundamentals of Deep Learning workshop serves as the primary entry point for higher education, requiring only basic Python knowledge.
- [04:00](https://www.youtube.com/watch?v=DIIteghPGx4&t=240s) The DLI Ambassador Program requires CV vetting, course completion, and a technical evaluation interview conducted by an NVIDIA master instructor.
- [04:00](https://www.youtube.com/watch?v=DIIteghPGx4&t=240s) NVIDIA supports certified ambassadors with up to $500 in catering reimbursements per workshop, travel expense coverage, and exclusive early access to new curricula.
- [07:25](https://www.youtube.com/watch?v=DIIteghPGx4&t=445s) Ambassador eligibility is strictly limited to faculty, permanent lecturers, and HPC center training staff, while graduate and PhD students are excluded due to turnover.
- [07:25](https://www.youtube.com/watch?v=DIIteghPGx4&t=445s) Ambassadors must teach at least two workshops every 12 months to maintain their active status, with tiered incentives offering financial rewards and GTC discounts.
- [07:25](https://www.youtube.com/watch?v=DIIteghPGx4&t=445s) Certified instructors can legally resell commercial workshop seats at market rates to private enterprise clients after buying them at educational discounts.
- [09:45](https://www.youtube.com/watch?v=DIIteghPGx4&t=585s) NVIDIA Teaching Kits provide a low-touch alternative to the Ambassador Program, delivering downloadable, full-semester instructional assets with no certification prerequisite.
- [10:12](https://www.youtube.com/watch?v=DIIteghPGx4&t=612s) Seven specialized Teaching Kits cover domains including deep learning, RAPIDS accelerated data science, Jetson edge AI, physics-informed Modulus simulations, and OpenUSD.
- [15:53](https://www.youtube.com/watch?v=DIIteghPGx4&t=953s) Official associate and professional certification exams administered through Certiverse represent the only paid tier within the DLI educational framework.
- [17:48](https://www.youtube.com/watch?v=DIIteghPGx4&t=1068s) ECPI University leverages DLI workshops and Teaching Kits across diverse disciplines, including computer science, robotics, and non-technical majors like culinary arts.
- [21:33](https://www.youtube.com/watch?v=DIIteghPGx4&t=1293s) Hands-on workshops require active code execution on cloud GPUs, scaling to over 100 students per session when supported by trained teaching assistants.
- [30:35](https://www.youtube.com/watch?v=DIIteghPGx4&t=1835s) Workshop delivery is flexible and can be divided into 60- to 90-minute modules to fit standard class periods, provided cloud GPU instances are properly shut down.
- [32:55](https://www.youtube.com/watch?v=DIIteghPGx4&t=1975s) Institutions can run co-curricular, 8-hour Friday workshops at zero cost to students, achieving strong engagement and approximately 70% completion pass rates.
- [34:39](https://www.youtube.com/watch?v=DIIteghPGx4&t=2079s) University staff members whose official job duties involve technical training and hosting workshops are eligible for the DLI Ambassador Program alongside traditional faculty.
## Detailed walkthrough
### [00:04](https://www.youtube.com/watch?v=DIIteghPGx4&t=4s) The Deep Learning Institute and Cloud-Powered Learning Platform
The Deep Learning Institute (DLI) centers its educational offerings on full-day, six- to eight-hour courses designed for both instructor-led workshops and online, self-paced study. All of this content is deployed through NVIDIA's proprietary DLI platform—a customized deployment of Open edX engineered specifically to connect students directly with cloud-hosted graphics processing units (GPUs). Rather than dealing with local hardware limitations, learners get real, interactive cloud GPU access to execute complex laboratory experiments and develop end-to-end software applications straight from their browsers.
To ensure students master the material, the courses rely on integrated assessments equipped with automated grading functionality. When a student successfully completes and passes these auto-graded labs, the system generates an official, individually named certificate. These credentials can be showcased on resumes and LinkedIn profiles, serving as tangible resume-building milestones for students entering the technical job market. When taken independently, all an aspiring learner needs is a standard laptop and a stable internet connection; when delivered live, classes are led by an NVIDIA-certified instructor who has undergone rigorous vetting.
The educational scope covers NVIDIA's hardware and software ecosystem. Offerings range from agentic AI, physical AI, generative AI, and large language models (LLMs) to specialized accelerated computing tracks. In the accelerated computing courses, students learn CUDA programming to speed up domain-specific scientific algorithms that often have nothing to do with machine learning, as well as edge computing architectures.
* **Course Formats:** 6- to 8-hour comprehensive modules delivered either live by certified instructors or as self-paced online modules.
* **Infrastructure:** Custom Open edX platform wired directly to cloud GPUs, eliminating local hardware bottlenecks.
* **Credentialing:** Verifiable, individualized completion certificates generated via embedded auto-graded code assessments.
* **Curriculum Breadth:** Agentic AI, physical AI, generative AI, LLMs, domain science acceleration with CUDA, and edge computing.
> “These courses run on what's called the DLI platform. This is an OpenEdX-based platform that we have customized to allow learners to get access to real GPUs in the cloud to run different lab experiments and develop applications like I like I said within these courses.”
---
[01:54](https://www.youtube.com/watch?v=DIIteghPGx4&t=114s) Workshop Catalog and the Flagship Deep Learning Fundamentals
The live instructor-led catalog organizes technical domains into dedicated tracks: Generative AI and LLMs, deep learning fundamentals, domain-specific application areas, data science, accelerated computing, and physical AI simulation. During major events such as the “Day Zero” Sunday sessions at GTC, NVIDIA runs as many as eight of these intensive workshops concurrently. A typical workshop unfolds across several sequential modules broken up by periodic intermissions, with the certified instructor fielding technical questions, debugging code, and steering students through intricate lab tasks in real time.
```mermaid
flowchart TD
accTitle: DLI Instructor-Led Workshop Tracks
accDescr: Diagram categorizing the six core tracks offered through NVIDIA DLI instructor-led workshops.
subgraph CoreAI["Foundational & Advanced AI"]
A["GenAI & LLMs"]
B["Fundamentals of Deep Learning"]
C["Domain-Specific Applications"]
end
CoreAI --> G["Cloud GPU-Accelerated Delivery"]
Systems --> G
```
Among all available curricula, the *Fundamentals of Deep Learning* course stands out as the undisputed staple within higher education. It serves as the primary gateway for academic faculty stepping into the DLI ecosystem. The barrier to entry for learners is intentionally approachable: students need only a foundational grasp of Python programming to succeed. Throughout the day, students get hands-on experience loading pre-trained models, working with large language models, and executing neural inference workflows. Crucially, academic departments do not have to purchase, configure, or maintain local lab servers to run these workshops because NVIDIA absorbs the computing overhead entirely in the cloud. Beyond introductory deep learning, faculty also leverage specialized modules such as diffusion models and full application development with large language models.
> “Educators using this course they don't need to configure anything in their labs. It's all it's all taken care of by by NVIDIA.”
---
[04:00](https://www.youtube.com/watch?v=DIIteghPGx4&t=240s) The DLI Ambassador Program and Instructor Certification
To scale these workshops into collegiate classrooms worldwide, NVIDIA manages dedicated educator programs split into two primary initiatives: the high-touch DLI Ambassador Program and the self-serve Teaching Kit Program. The Ambassador Program is built around a formal instructor certification pipeline that functions as an application-based vetting process. Interested university educators must first demonstrate relevant technical domain experience. NVIDIA then grants them student-level access to the target workshop on the DLI platform. Candidates must work through the syllabus, pass all auto-graded assessments, and conclude the process with a technical evaluation interview conducted by an NVIDIA master instructor.
Once faculty pass the interview and achieve certification, they are licensed to deliver the workshop across any academic environment at no charge. Ambassadors frequently incorporate these workshops directly into their standard semester syllabi, host free standalone campus bootcamps, or present them as official conference tutorials at venues like the annual Supercomputing (SC) conference. The primary incentive for faculty is immediate access to NVIDIA’s freshest technical content. Newly authored workshops almost always debut exclusively in an instructor-led format long before any self-paced version is released, meaning universities must sponsor a certified ambassador to bring the newest curricula to their students.
While the institutional prestige of holding an NVIDIA teaching credential is a common draw, the speaker acknowledges that it can backfire if faculty view it solely as a resume ornament: some professors complete the process to collect the badge on LinkedIn but never actually teach, which eventually leads to their credentials expiring. For active educators, NVIDIA sweetens the program by reimbursing up to $500 per workshop for event catering—such as food and drinks to draw student attendance—and covering travel expenses when instructors travel to teach at peer universities or conferences. The global ambassador cohort is deliberately kept close-knit, with just under a thousand active members worldwide who regularly advise NVIDIA on curriculum development and academic exam design.
```mermaid
flowchart LR
accTitle: DLI Ambassador Certification and Delivery Pipeline
accDescr: Four-stage pathway showing how faculty apply, get certified, and deliver DLI workshops.
A["Faculty Application & CV Review"] --> B["Self-Paced Course Completion"]
B --> C["Pass Technical Interview"]
C --> D["Certified DLI Ambassador"]
D --> E["Teach On Campus"]
D --> F["Conference Tutorials (e.g., SC)"]
D --> G["Discounted Enterprise Delivery"]
```
> “Anytime that we seek guidance or want to learn about academia and teaching and training we go to our ambassadors first and foremost and they help us develop content, develop certification exams that I'll talk about.”
---
[07:25](https://www.youtube.com/watch?v=DIIteghPGx4&t=445s) Tiered Perks, Enterprise Teaching, and Faculty Eligibility
While the Ambassador Program is completely free for academic instruction, NVIDIA also allows certified faculty to monetize their training skills in the private sector on a modest scale. Ambassadors who have ties to enterprise clients can purchase commercial workshop seats from NVIDIA at a steep educational discount and resell the training at market rate to corporate customers, turning a direct profit for their lab or personal consulting practice.
Within the academic track, NVIDIA maintains an incentive structure modeled after an airline frequent-flyer program. As ambassadors deliver more workshops and graduate larger volumes of certified students, they ascend through recognized achievement tiers. Reaching higher tiers unlocks tangible rewards, including direct financial payments, discounted registration for the annual GTC conference, and enhanced public profiles on NVIDIA’s official online Instructor Directory.
Maintaining this high-touch program requires strict eligibility guardrails. To be accepted, applicants must hold a role as university faculty, full-time lecturers, or technical staff directly responsible for training at institutional High-Performance Computing (HPC) centers. Although NVIDIA receives a heavy volume of applications from undergraduate and PhD students, these are routinely declined. The program administrators found that graduate students cycle out quickly upon finishing their degrees, whereas permanent faculty remain at their institutions for years and sustain consistent, long-term teaching. Once certified, ambassadors must teach a minimum of two workshops every 12 months to maintain their active standing in the directory.
* **Commercial Resale Rights:** Ambassadors can purchase corporate workshop seats at heavy discounts and resell instruction to enterprise clients.
* **Frequent-Flyer Tiers:** Higher workshop volume unlocks financial stipends, GTC registration discounts, and directory honors.
* **Eligible Roles:** Faculty, full-time academic lecturers, and HPC center training leads only.
* **Student Ineligibility:** PhD candidates and students are excluded due to program capacity and high post-graduation turnover.
* **Activity Requirement:** Must conduct at least two workshops per 12-month cycle to avoid certification expiration.
> “We do get a lot of a lot of applications from students and PhD students and um and we have to decline those for for a number of different reasons. We just get way too many applications. Um, and the the faculty tend to stick around and and actually do do a lot more teaching over time.”
---
[09:45](https://www.youtube.com/watch?v=DIIteghPGx4&t=585s) Transitioning to Teaching Kits: A Low-Touch Resource for Faculty
For professors seeking instructional materials without committing to the rigorous vetting and interview process of the Ambassador Program, the speaker points to the Teaching Kits program as the ideal entry point. Described as a low-touch alternative, the Teaching Kits are hosted on an accessible download portal. Once verified, educators can directly download massive, pre-packaged digital repositories containing complete, semester-long instructional curricula designed for immediate integration into university degree programs.
> “The program that I recommend everybody get started with is the teaching kits program. This is a much easier access lower touch program for us and what this looks like is effectively a portal that educators access and they within this portal they can just download these very large packages of semester long teaching materials.”
### [10:12](https://www.youtube.com/watch?v=DIIteghPGx4&t=612s) Comprehensive Teaching Kits for Educators
The speaker details a library of seven comprehensive teaching kits designed for university faculty. Unlike shorter, supplemental resources, these kits are intended to provide full-semester content that can be downloaded and adapted. Each package is built to be modular, allowing educators to "pick and choose" specific modules or build entire courses from scratch. These resources, which include lecture slides, videos, Jupyter notebooks running in free Colab instances, textbooks, and problem sets, are distributed under a non-commercial Creative Commons license, providing faculty the flexibility to modify the content as needed.
NVIDIA emphasizes that these materials are co-developed with academics to ensure a balance between theoretical depth—such as the math behind deep neural networks—and the practical, industry-standard application focus that students need. Furthermore, the kits are integrated with NVIDIA’s Deep Learning Institute (DLI) platform. The self-paced courses, which offer student assessments and certifications, function as a subset of these larger teaching kits.
* **Key Teaching Kits:**
* **Deep Learning:** Heavily influenced by Yann LeCun’s NYU course; focuses on the theory and mathematics of training deep neural networks.
* **Accelerated Data Science:** Developed in partnership with Georgia Tech and Prairie View A&M; centers on the RAPIDS open-source framework.
* **Edge AI and Robotics:** Based on the NVIDIA Jetson Nano; currently undergoing a major update for a fall release.
* **Deep Learning for Science and Engineering:** Utilizes the Modulus (referred to as "physics Nemo") platform for applications like computational fluid dynamics and biomedicine.
* **Generative AI:** Developed with Professor Sam Raman from Dartmouth; includes extensive exam prep aligned with NVIDIA certification exams.
* **Open USD:** Announced at GTC, covering 3D graphics, simulation, and physical AI; aligned with a new certification exam.
### [15:53](https://www.youtube.com/watch?v=DIIteghPGx4&t=953s) Certification Exams and Professional Pathways
NVIDIA offers proctored, online certification exams at both the associate and professional levels. These are the only components of the program that carry a cost, as they are administered through a third-party partner, Certiverse. The goal is to provide a standardized, recognized qualification that makes students more competitive in the job market. During GTC, these exams were offered for free to attendees at the Signia Hilton Hotel.
The presenter highlights the two primary pathways for faculty to get involved:
* **Teaching Kits:** Accessible via a simple 5-minute application form with typically rapid approval.
* **Ambassador Program:** A more involved process that requires a resume-style document demonstrating the educator's background, intended to certify them to teach specific workshops to their students.
### [17:48](https://www.youtube.com/watch?v=DIIteghPGx4&t=1068s) Implementing AI Training at ECPI University
Professor Paul Nussbaum of ECPI University describes how his institution leverages these NVIDIA programs to address student anxiety regarding career prospects. As an accelerated, career-focused university with campuses across the Mid-Atlantic, Florida, and Texas, ECPI views AI training as a critical component of its curriculum. They emphasize a "holistic" approach, teaching students not just how to create AI solutions, but how to use them across various fields—even including non-technical areas like culinary arts.
"We feel that it's critical for our students to really be prepared for a workforce where they may be creating these AI solutions or they may just be using those AI solutions," says Professor Nussbaum.
ECPI’s strategy integrates these NVIDIA resources into their existing computer science and electronics engineering technology programs. The focus is specifically on preparing graduates for high-growth sectors such as robotics and manufacturing automation, particularly as industrial facilities are increasingly "onshored" back to the United States.
```mermaid
flowchart TD
A["ECPI Strategy"] --> B["Academic Programs"]
A --> C["Hands-on Workshops"]
A --> D["Interdisciplinary Integration"]
B --> E["Computer Science & Engineering"]
C --> F["NVIDIA DLI Certifications"]
D --> G["AI in Culinary & Other Fields"]
accTitle: ECPI University AI Training Strategy
accDescr: A flowchart showing how ECPI University uses a multi-faceted strategy to integrate AI, ranging from academic coursework and NVIDIA partnerships to interdisciplinary applications.
```
### [21:33](https://www.youtube.com/watch?v=DIIteghPGx4&t=1293s) The Ambassador Experience and Student Engagement
Paul discusses his role as an NVIDIA ambassador, emphasizing that the eight-hour workshops he hosts are voluntary extracurricular activities. Despite requiring students to dedicate significant personal time outside of their regular coursework, these sessions are highly sought after. Paul maintains a capacity of 20 attendees per workshop but can scale this up by 20 for every teaching assistant present, often reaching 100 students per session. He notes that the NVIDIA brand is a major draw, respected by students not only for its prominence in gaming and personal hardware but also for the prestige and legitimacy of its certifications.
“I know I'm going to get a workshop where you cannot cheat to pass the assessment,” Paul explains, highlighting why he trusts the program's rigor and value for his students’ portfolios and LinkedIn profiles. Beyond the brand, he emphasizes the hands-on nature of the training: “I tell students, this is not a lean-back webinar that you attend for 8 hours. This is a lean-forwards hands-on.” Students actively modify code and train models using cloud-based GPUs, which Paul jokingly compares to “driving a Ferrari” compared to the slower processing speeds of their personal laptops.
### [25:02](https://www.youtube.com/watch?v=DIIteghPGx4&t=1502s) Curriculum and Workshop Variety
The workshop catalog covers a wide range of specialized topics designed to meet different student interests. Anomaly detection is particularly popular with cybersecurity students due to its focus on processing massive log files to identify threats. Other offerings include:
* **Industrial Automation:** Predictive maintenance workshops, currently being restructured to encompass a broader scope.
* **Generative AI:** Diffusion models, which teach students the underlying mechanics of text-to-image generation—a topic that satisfies the “hungry curiosity” of electronics engineering students.
* **Jetson Workshops:** A more accessible entry point that does not require Python proficiency, focusing instead on data collection and neural network training via transfer learning.
Paul illustrates the importance of robust training data with an anecdote about a Jetson workshop where students trained a model to recognize hand gestures. By wearing a distinctively patterned shirt that was intentionally excluded from the initial training set, Paul demonstrates to the students how model accuracy shifts when encountering data variations, reinforcing key machine learning concepts through practical, often humorous, interaction.
### [28:10](https://www.youtube.com/watch?v=DIIteghPGx4&t=1690s) Certification, Interdisciplinary Integration, and Future Goals
The program is increasingly focusing on structured certification paths, supported by a newly launched portal that helps students track prerequisites and prepare for exams. ECPI University also emphasizes interdisciplinary projects, such as the “Chef’s Buddy” initiative—an AI-driven recipe and image generation tool that brings together students with varying levels of AI knowledge. According to Paul, this is part of a broader strategy to weave AI into the fabric of the entire university, including business degree programs, to ensure students are “ready to hit the ground running” when they enter the workforce.
Looking ahead, Paul is preparing to teach new workshops on Agentic AI, specifically focusing on Retrieval-Augmented Generation (RAG) and AI agent development. He encourages other faculty members to pursue the ambassador program, characterizing it as a significant boost to his own professional development and a transformative opportunity for his students.
### [30:35](https://www.youtube.com/watch?v=DIIteghPGx4&t=1835s) Q&A: Workshop Flexibility and Academic Integration
During the Q&A session, audience members sought clarification on logistics and integration. One attendee asked if the eight-hour assessments must be completed in one sitting. The speakers clarified that they provide significant flexibility; the workshops can be broken down into shorter segments, such as 60-to-90-minute sessions, to fit within standard class meeting times. However, they reminded faculty to ensure students properly shut down their GPU instances between sessions to avoid unnecessary cloud resource consumption.
When asked by a representative from Hong Kong Polytechnic University about the difficulties of embedding these materials into formal academic subjects, the speakers pointed to the NVIDIA teaching kits as the primary solution. These kits are designed to be modular, allowing professors to “pick and choose” specific modules to incorporate into their existing syllabi, effectively lowering the barrier for integration into university curricula.
[32:55](https://www.youtube.com/watch?v=DIIteghPGx4&t=1975s) Offering Friday Co-Curricular Workshops and Student Certification
An audience member asks how the workshops are integrated into student schedules, prompting one of the presenting faculty members to share how they structure the sessions outside normal class hours. At their institution, regular classes run Monday through Thursday, so workshops are offered as co-curricular Friday sessions. The pitch to students is straightforward and compelling: attend for an intensive eight-hour day and walk away with a numbered, traceable certificate from NVIDIA. Because many of these students are non-traditional learners juggling families, careers, and heavy workloads, having a structured, single-day opportunity to earn a verifiable industry credential is a huge draw.
The presenter notes that while not every participant passes—the workshops maintain roughly a 70% pass rate—the engagement and educational value remain exceptionally high across the board. Even participants who do not intend to become career Python programmers enjoy the material and gain a practical understanding of foundational artificial intelligence concepts. When an attendee inquires about the logistics, asking about major startup challenges and whether students or academic departments are charged for the additional resources, the presenter clarifies that there is absolutely no fee charged to the students.
The presenter contributes their own teaching time without extra compensation, treating the effort as part of their service and professional advancement. While teaching assistants occasionally receive small tokens of appreciation such as gift cards, the faculty time is entirely donated. The questioner emphasizes that understanding these resource constraints is critical if an institution wants to scale up and mobilize multiple faculty members or instructors to adopt the same model, and the presenter welcomes further one-on-one discussion after the session to unpack the practical operational details.
* “We say earn a numbered and traceable certificate from Nvidia in eight hours.”
* “I've got about a 70% pass rate on many of these workshops but they all learn and they all love it even if they don't necessarily want to become a Python programmer.”
* “There's no fee to the students.”
---
[34:39](https://www.youtube.com/watch?v=DIIteghPGx4&t=2079s) University Staff Eligibility for the NVIDIA Ambassador Program
A staff member from the University of Texas at Austin raises a question regarding the NVIDIA Deep Learning Institute (DLI) University Ambassador Program. Noting that the presentation slides explicitly highlighted university faculty, the attendee asks whether university staff members are also eligible to become certified ambassadors.
The NVIDIA representative clarifies that university staff members can qualify, provided their day-to-day responsibilities align with the program's educational goals. The primary evaluation criteria center on employment status and whether the applicant's regular institutional role involves running workshops, delivering technical training, and teaching hands-on material. If an applicant has consistent access to relevant academic audiences and regular opportunities to lead training sessions, staff members are welcomed into the ambassador pipeline. Satisfied with the clarification, the UT Austin staff member jokingly notes that they forgot their planned second question, concluding their turn at the microphone.
* “It is especially if you are as part of your job training and teaching and running workshops.”
* “The real requirement is a level of employment. So again then from there we would sort of look at what you're doing on a daily basis and do we think that you're going to be teaching and have the opportunity to teach workshops and have access to the audiences and all of that.”
---
[35:32](https://www.youtube.com/watch?v=DIIteghPGx4&t=2132s) Cross-Disciplinary AI Minors in the Dominican Republic
With time running out, the session takes one final inquiry from the CEO of an artificial intelligence and robotics innovation lab based in the Dominican Republic. The speaker explains that their lab advises two of the leading universities in the Dominican Republic and has designed an applied minor in AI and emerging technologies intended for students across every professional field. Rather than restricting AI education to computer science or engineering departments, the program is tailored so that aspiring lawyers, physicians, and other non-technical professionals complete an applied AI curriculum focused directly on their discipline.
```mermaid
flowchart TD
accTitle: Cross-Disciplinary AI Minor Structure
accDescr: Diagram showing how a general university degree feeds into an applied AI minor tailored for specific professional paths.
subgraph AIMinor ["Final Semester: Applied AI Minor"]
M["AI & Emerging Technologies Minor"]
L["Applied AI for Law"]
D["Applied AI for Healthcare & Medicine"]
O["Applied AI for Other Careers"]
end
U --> M
M --> L
M --> D
M --> O
```
The minor is structured as a dedicated final semester added to the conclusion of the students' degree programs, and it is scheduled to officially launch immediately that coming weekend. Because the innovation lab works in direct coordination with the university rectors and academic leadership of both institutions, the speaker emphasizes that incorporating NVIDIA's instructional roadmaps, teaching kits, and certification workshops into this newly established program would be seamless. They express a strong desire to explore formal collaboration with NVIDIA to integrate its accelerated computing and AI curriculum into this nationwide cross-disciplinary initiative.
* “We minor in AI and emerging technologies applied to every career. So the lawyer will be out with an minor with AI applied to the lawyer the professional and the doctor and everyone.”
* “This is one more semester at the end of all the program.”
* “I have contact with the rect with the the principal of the of the both universities so it will be very easy to me to incorporate all of all of these to that program.”
## Notable quotes
> [21:33](https://www.youtube.com/watch?v=DIIteghPGx4&t=1293s) I tell students, this is not a lean-back webinar that you attend for 8 hours. This is a lean-forwards hands-on. — Paul Nussbaum
> [17:48](https://www.youtube.com/watch?v=DIIteghPGx4&t=1068s) We feel that it's critical for our students to really be prepared for a workforce where they may be creating these AI solutions or they may just be using those AI solutions — Paul Nussbaum
## Mentioned
NVIDIA, Deep Learning Institute (DLI), Open edX, CUDA, GTC, Supercomputing (SC), DLI Ambassador Program, Teaching Kit Program, Yann LeCun, New York University (NYU), RAPIDS, Georgia Tech, Prairie View A&M University, NVIDIA Jetson Nano, Modulus, Sam Raman, Dartmouth College, OpenUSD, Certiverse, ECPI University, Paul Nussbaum, Lebanese American University, Hong Kong Polytechnic University, University of Texas at Austin
## Video description
NVIDIA’s developer training programs support faculty teaching students and researchers through downloadable teaching kits, a cloud-based GPU-accelerated platform, and technical certification exams. Instructional content spans foundational topics—deep learning, accelerated computing, and data science—and advanced areas including agentic AI, natural language processing, and generative AI. Online courses and instructor-led workshops with project-based assessments help faculty align coursework with certifications and support students in earning recognized qualifications. We highlight ECPI integrating GPU-accelerated AI into student projects, and Lebanese American University using hands-on, faculty-led workshops to advance applied learning and prepare students for real-world AI applications and industry pathways.
NVIDIA's Deep Learning Institute provides higher education institutions with cloud-hosted graphics processing unit infrastructure, modular curricula, and educator training pipelines to scale artificial intelligence and high-performance computing instruction without local hardware expenses.
Cloud Infrastructure and Automated Assessment Pipeline (0:04)
The Deep Learning Institute (DLI) delivers full-day, six- to eight-hour courses through a customized Open edX platform wired directly to cloud GPUs. This setup allows students to run laboratory experiments and build end-to-end applications within standard web browsers, eliminating the need for academic departments to purchase, configure, or maintain local computing clusters.
The educational scope spans large language models (LLMs), agentic AI, physical AI simulation, edge architectures, and scientific algorithm acceleration using CUDA. Rather than passive lectures, workshops use embedded coding assessments with automated grading. When students successfully pass these integrated labs, the platform issues an individualized, verifiable certificate that learners can display on resumes and LinkedIn profiles.
Dual Enablement Models: Ambassadors and Teaching Kits (4:00)
NVIDIA segments university support into two operational tiers: low-touch Teaching Kits and the high-touch DLI Ambassador Program. Teaching Kits provide open, downloadable semester-long curricula distributed under non-commercial Creative Commons licenses. Developed alongside academic partners like NYU, Georgia Tech, and Dartmouth, these packages supply lecture slides, Colab-compatible Jupyter notebooks, problem sets, and textbooks across subjects such as accelerated data science, robotics, generative AI, and OpenUSD.
The Ambassador Program trains and licenses full-time faculty, lecturers, and institutional high-performance computing staff to deliver live DLI workshops for free. Candidates must pass the target course and a technical interview with an NVIDIA master instructor. Once certified, ambassadors gain exclusive access to early-release workshops, receive up to $500 per event in catering reimbursements, and can access travel funding. To ensure instructional longevity, undergraduate and doctoral students are excluded from the ambassador track, and active instructors must deliver at least two workshops per year.
Institutional Adoption and Student Credentials at ECPI (17:48)
Universities apply these training resources to connect classroom instruction with career readiness. At ECPI University, Professor Paul Nussbaum delivers voluntary, eight-hour Friday bootcamps that scale up to 100 students when supported by teaching assistants. Because the courses target non-traditional students balancing employment and family responsibilities, earning a verifiable industry credential in a single day acts as a strong incentive.
Nussbaum noted an approximate 70% student pass rate on the auto-graded assessments, describing the curriculum as an active, hands-on environment where students cannot bypass rigor. Workshop topics cover anomaly detection in cybersecurity log files, generative diffusion models, and computer vision on Jetson hardware. ECPI also integrates these tools into non-technical degree programs, including business and culinary arts initiatives like the "Chef's Buddy" recipe generation tool.
Curriculum Modularity, Certification, and Program Expansion (30:35)
To accommodate conventional university timetables, NVIDIA allows instructors to split the standard eight-hour workshops into 60-to-90-minute classroom segments, provided cloud GPU instances are powered down between meetings to conserve resources. While workshop attendance and course materials are free to academic institutions, formal proctored certifications administered online by Certiverse at the associate and professional levels represent the only paid offering in the ecosystem.
Institutional adoption continues to broaden beyond tenured professors. Program administrators confirmed that university staff actively engaged in campus technical training, such as high-performance computing support personnel, are eligible for ambassador status. Additionally, international universities, including partners in the Dominican Republic, are embedding DLI modules into cross-disciplinary emerging technology minors required across professional fields like law and medicine.
NVIDIA's enablement programs remove compute and curriculum barriers for universities, yet their delivery model relies on volunteer faculty time and active monitoring of cloud resource consumption. While workshop attendance remains free, academic departments must navigate the distinction between internal completion certificates and paid, third-party proctored certification exams for formal workforce credentials.
Source quotes
[21:33] Paul Nussbaum: I tell students, this is not a lean-back webinar that you attend for 8 hours. This is a lean-forwards hands-on.
[17:48] Paul Nussbaum: We feel that it's critical for our students to really be prepared for a workforce where they may be creating these AI solutions or they may just be using those AI solutions
Original transcript
[0:04] Um, so
[0:06] if you if you look at the Deep Learning
[0:08] Institute, what the what is mostly
[0:10] offered are 6 to 8-hour
[0:13] full-day courses. These are taught
[0:15] instructor-led and both self online
[0:18] self-paced as well.
[0:20] Um, these courses run on what's called
[0:23] the DLI platform. This is an
[0:25] OpenEdX-based platform that we have
[0:27] customized to allow learners to get
[0:29] access to real GPUs in the cloud to run
[0:34] different lab experiments and develop
[0:36] applications like I like I said within
[0:38] these courses.
[0:40] Um, and they're based on assessments
[0:42] built into the course that have the an
[0:44] auto grading functionality to them and
[0:46] if the learners pass the assessments
[0:48] then they get a unique certificate um,
[0:50] that has their name on it and they can
[0:52] put these certificates on their
[0:53] LinkedIn, on the resume and uh it's a
[0:56] real a really good uh career building
[0:58] exercise for for the students.
[1:00] Um, these courses
[1:03] uh again uh can be done instructor-led.
[1:06] Uh these are taught by a certified
[1:07] instructor that's gone through a
[1:10] certification process with NVIDIA.
[1:12] Um, or again they can be done completely
[1:15] self-paced where all you need is a
[1:17] laptop and a reliable internet
[1:18] connection.
[1:19] >> [snorts]
[1:21] [clears throat]
[1:22] >> And these courses and workshops uh cover
[1:25] really all areas of technology related
[1:28] to NVIDIA um, from agentic AI to
[1:31] physical AI,
[1:33] um, generative AI and large language
[1:35] models. We also have um, accelerated
[1:37] computing courses where you're using
[1:39] CUDA to accelerate domain science types
[1:42] of algorithms that maybe not don't have
[1:44] anything to do with AI,
[1:46] um, edge computing. So we have um,
[1:48] really a a breadth of
[1:50] uh courses that again cover all these
[1:52] technologies.
[1:54] This is a look at the current
[1:57] instructor-led workshops. So we had we
[1:59] ran eight of these workshops on Sunday
[2:03] sort of the day zero of GTC.
[2:06] Um, I don't know if any of you guys were
[2:07] able to attend.
[2:09] Um, but this is again sort of the
[2:10] catalog of instructor-led workshops and
[2:12] you can see how we how we divide these
[2:14] workshops between GenAI and LLM, deep
[2:17] learning and and fundamentals of deep
[2:19] learning and then some of the courses go
[2:21] into more specific application domains
[2:24] uh data science, accelerated computing
[2:26] and then also simulation and physical
[2:28] AI. So these are the
[2:30] the workshops that are offered
[2:31] instructor-led.
[2:33] And what a instructor-led workshop
[2:35] typically looks like is you have a
[2:38] number of sort of modules
[2:40] within the day and a few breaks in
[2:42] between. Of course the benefit of
[2:45] attending an instructor-led workshop is
[2:47] you get access to a certified instructor
[2:49] to ask question questions and help you
[2:51] guide through through the course.
[2:54] Um [snorts]
[2:56] and this is an example course outline.
[2:58] So the fundamentals of deep learning is
[3:00] by far the most popular course in higher
[3:03] ed today.
[3:04] This is where a lot of instructors in
[3:06] academia um start in terms of teaching
[3:09] workshops under the Deep Learning
[3:11] Institute. And really students only need
[3:14] a basic background in Python programming
[3:16] to be pretty successful in this
[3:18] workshop. Uh so they they get hands-on
[3:22] um to do some some interesting things in
[3:24] terms of uh using pre-trained models and
[3:27] large language models um
[3:29] and then doing some some inference type
[3:31] of type of things. And again this is all
[3:32] running on real GPUs in the cloud on the
[3:34] back end. Um, educators using this
[3:37] course they don't need to configure
[3:38] anything uh in their labs. It's all it's
[3:41] all taken care of by by NVIDIA.
[3:44] And we have some more advanced workshops
[3:46] in things like diffusion models. That's
[3:48] a also a really popular course.
[3:50] Um, also um, application development
[3:53] using large language models. So I'm not
[3:54] going to go into the details of of these
[3:56] courses but again lots and lots of
[3:58] instructor-led workshops.
[4:00] >> [snorts]
[4:00] >> And I'm kind of going through through
[4:02] these offerings to let you know that
[4:05] uh through our educator programs which I
[4:07] primarily manage
[4:09] um,
[4:09] we enable faculty at universities to get
[4:13] access to these workshops and to get
[4:15] access to a ton of other content that
[4:17] I'll talk about in a second.
[4:18] Um, so the DLI ambassador program if
[4:20] you've ever heard of this I'm going to
[4:21] start with this program. So we have two
[4:23] distinct programs, the ambassador
[4:25] program and the teaching kit program.
[4:27] Um, the ambassador program is a bit
[4:29] higher touch program for us. Um, it
[4:32] starts with a instructor certification
[4:35] process. We call it instructor
[4:36] certification. It's really kind of a
[4:38] application-based vetting process. Um,
[4:41] so an an educator who is interested in
[4:43] in getting certified to teach any one or
[4:46] more of those workshops would would fill
[4:48] out an application.
[4:50] Um, they would need to show us that they
[4:51] have some uh relevant technical
[4:53] background in the workshops they're
[4:55] they're seeking to get certified to
[4:57] teach. As part of the process we'd give
[4:59] them access to that course. They would
[5:01] take it online as a student would take
[5:03] it, uh complete the assessments and this
[5:06] process would culminate with an
[5:07] interview
[5:08] uh with a instructor on NVIDIA's side
[5:12] who has the expertise to certify other
[5:13] instructors in that particular workshop.
[5:16] >> [snorts]
[5:17] >> And if we pass the interview then you
[5:19] get access to teach that course
[5:21] effectively in any academic setting and
[5:24] uh you know you could teach the course
[5:26] to students in your in your class or to
[5:28] other university students. Um, you can
[5:31] bring these to academic conferences.
[5:33] We've seen these workshops being used as
[5:35] tutorials at conferences like
[5:37] supercomputing.
[5:38] Um
[5:40] yeah. And so why would anyone want to go
[5:42] through this process? Well, um, it's
[5:46] first and foremost I think the
[5:47] motivations are the content, right? So
[5:49] this is sort of cutting-edge world-class
[5:51] content coming from NVIDIA. Um, anytime
[5:55] the a new workshop becomes available it
[5:58] usually starts only as an instructor-led
[6:00] workshop not a self-paced version. So in
[6:03] order to get access to the latest
[6:05] workshops uh universities usually need
[6:07] to get certified to be able to teach it
[6:09] on on their campus.
[6:11] Um, there's some other motivating uh
[6:13] factors here. So uh some folks are
[6:15] motivated by the credential of being an
[6:17] instructor-led uh sorry being a
[6:19] certified instructor with with NVIDIA.
[6:21] Um, you know they get their own uh
[6:24] unique certificates that have their name
[6:25] on it. They put that on their LinkedIn.
[6:28] Uh sometimes um it's a problem that that
[6:30] is the only motivating factor. So they
[6:32] come in and they get their instructor
[6:33] certifications but never teach.
[6:35] Um, so then they eventually sort of
[6:36] expire out of the program.
[6:38] >> [snorts]
[6:38] >> Um, other other perks to the program. So
[6:40] if you guys want to hold a workshop and
[6:43] cater the event, bring [clears throat]
[6:44] in some food and beverage uh to bring
[6:46] students in we'll reimburse
[6:48] um up to $500 for workshop. We also
[6:50] reimburse for travel expenses if you
[6:52] want to teach at other universities or
[6:54] or conferences.
[6:56] Um, it's a it's a it's a pretty
[6:58] tight-knit community. So today we only
[7:00] have a a little less than a thousand
[7:03] ambassadors um,
[7:05] in in the world and uh and and we set up
[7:08] some uh networking opportunities with
[7:10] them. Anytime that we seek guidance or
[7:13] want to learn about academia and
[7:16] teaching and training we go to our
[7:17] ambassadors first and foremost and uh
[7:19] they help us [clears throat] develop
[7:20] content, develop certification exams
[7:22] that I'll talk about.
[7:24] Um
[7:25] yeah and so another benefit I'll
[7:27] mention. So again this this program is
[7:29] all free for use in academia. Um, some
[7:33] of our ambassadors have access to
[7:35] enterprise customers who are looking for
[7:37] training as well. So on a relatively
[7:38] small scale our ambassadors can also
[7:41] teach workshops to enterprise and they
[7:44] can effectively buy the workshop that
[7:46] they're certified to teach from us at a
[7:48] a very good discount and then they can
[7:50] resell that to their enterprise
[7:51] customers um to turn a profit if they
[7:54] want to.
[7:57] Uh some additional benefits we kind of
[7:59] have this frequent flyer type of program
[8:01] where
[8:02] um the more workshops you're delivering,
[8:04] the more students that are getting
[8:05] certificates in the workshops it starts
[8:08] to unlock additional benefits.
[8:10] Um,
[8:11] benefits include actual financial
[8:13] payments. Again this this motivates some
[8:15] more than others. Um,
[8:17] and uh but other things like access to
[8:20] GTC conference discounts, registration
[8:23] discounts. Um
[8:24] and yeah so so the ambassadors
[8:26] effectively start reaching these these
[8:28] tiers. Uh a lot of them are very proud
[8:30] of these tiers.
[8:32] Um, we have a a formal instructor
[8:35] directory which is kind of the external
[8:37] validation that they have these
[8:38] credentials and all of these these extra
[8:41] things show up there as well.
[8:44] >> [snorts]
[8:44] >> Uh in terms of the eligibility um so we
[8:46] do require that ambassadors are
[8:49] um faculty or full-time lecturers.
[8:52] Uh we also certify those folks who are
[8:55] responsible for training at their HPC
[8:57] centers. Um, but really you know
[9:00] for the most part we do get a lot of a
[9:02] lot of applications from students and
[9:04] PhD students and um and we have to
[9:06] decline those for for a number of
[9:08] different reasons. We just get way too
[9:09] many applications. Um, and the the
[9:12] faculty tend to stick around and and
[9:13] actually do do a lot more teaching over
[9:15] time.
[9:16] And [snorts] again there are technical
[9:18] qualifications for each workshop that uh
[9:21] that they apply for that we we check
[9:23] against to determine if we think they'll
[9:25] be successful and even put them through
[9:27] the process at all.
[9:29] Oh and then we we ask that our
[9:30] ambassadors teach at least a couple of
[9:32] workshops
[9:33] every 12 months to to stay active in the
[9:35] program.
[9:38] Uh and today again I mentioned we have
[9:40] um close to a thousand ambassadors today
[9:41] at at a number of different institutions
[9:43] across the world.
[9:45] Um, okay so that's the ambassador
[9:46] program. Um, the program that I
[9:48] recommend every everybody get started
[9:50] with is the teaching kits program. This
[9:52] is a much easier access lower touch
[9:55] program uh for us and what this looks
[9:58] like is effectively a portal that
[10:03] educators access and they within this
[10:06] portal they can just download these very
[10:07] large packages of semester long teaching
[10:10] materials. So pretty much everything
[10:12] I've talked about up until up until this
[10:14] point
[10:15] is pieces of content that's no longer
[10:17] than 6 to 8 hours. This is full semester
[10:21] level content that can be freely
[10:24] downloadable. These packages include
[10:27] typical teaching materials like lecture
[10:29] slides, lecture videos,
[10:32] Jupiter notebooks that have built been
[10:34] built out in free collab instances.
[10:38] Some of them have associated textbooks,
[10:39] Q&A type of problem sets.
[10:41] And and so you can see that we have
[10:44] seven teaching kits today and these
[10:47] cover sort of these broad broad areas of
[10:50] technology within Nvidia. And you can
[10:52] also see that we co-developed this
[10:53] content with other academics
[10:56] to to really sort of bridge that gap
[10:59] between industry standard application
[11:02] focused material with the the theory and
[11:05] sort of the math behind
[11:08] behind some of these things that that a
[11:10] lot of folks in academia do want to
[11:11] teach.
[11:14] So yeah, definitely check out the
[11:15] teaching teaching kits. Another aspect
[11:18] of the kits is that we weave in the
[11:21] self-paced versions of the courses that
[11:24] I talked about earlier and those courses
[11:26] also run on the DLI platform. So the 6
[11:30] to 8 hour self-paced courses that offer
[11:32] the students assessments are sort of a
[11:34] subset of these teaching kits. So so
[11:36] again, this is sort of where you want to
[11:39] start
[11:40] and then again, within that you get
[11:42] access to all different sorts of
[11:43] content.
[11:45] And you can see we here at GTC this week
[11:48] we're launching the newest teaching kit
[11:50] around open USD for 3D graphics and
[11:52] simulation and and physical AI and I'll
[11:54] talk a little bit about that
[11:56] in more detail. The next slides just
[11:58] sort of cover a little bit more into the
[12:01] types of content within each kit
[12:04] and sort of how how the content is
[12:06] structured.
[12:08] We build these these kits very
[12:10] modularly. So most educators are kind of
[12:13] coming in and picking and choosing which
[12:15] modules they want to introduce into
[12:17] their courses. Some are building their
[12:19] courses from scratch completely based on
[12:22] these teaching kits
[12:24] and some are just again coming in and
[12:25] just handpicking what they want to
[12:27] incorporate. It all comes under the
[12:30] the non-commercial creative commons
[12:32] license so you guys can freely modify
[12:35] this content
[12:36] as as you see fit to fit into your
[12:39] courses.
[12:40] The deep learning teaching kit, this was
[12:43] the second kit that we built
[12:45] and this was largely based on Jan
[12:47] LeCun's course at NYU. So this one goes
[12:50] pretty pretty heavy into the
[12:53] the theory and the math behind training
[12:55] deep neural networks.
[12:57] And you know, you know, this the core
[12:59] content here makes a good complement to
[13:02] the DLI courses that run on the platform
[13:04] again that are very hands-on, very
[13:07] application focused.
[13:09] >> [clears throat]
[13:10] >> Accelerated data science,
[13:12] we started this one about 4 years ago
[13:14] and it's been through a couple of
[13:15] updates.
[13:16] Built this one with professors at
[13:18] Georgia Tech and Prairie View A&M. So
[13:20] this one's largely based on the RAPIDS
[13:22] open source framework for accelerating
[13:25] more traditional machine learning
[13:27] algorithms and all of the things that
[13:30] have to do with data science.
[13:32] So definitely check that one out if if
[13:34] if you're interested in that.
[13:36] We have a kit around edge AI and
[13:38] robotics. This one's based on our
[13:39] embedded systems platform the Jetson or
[13:42] Nano.
[13:43] And we're doing a we're in the process
[13:46] of doing a big update to this one that
[13:48] we're going to release in the fall.
[13:52] So yeah, we have also one around deep
[13:55] learning for science and engineering.
[13:58] So this is using the physics Nemo
[14:01] platform. It's a physics ML based
[14:03] platform for accelerating all types of
[14:06] things within
[14:08] science domains like
[14:11] computational fluid and and and
[14:12] biomedicine
[14:14] and all these
[14:15] again, really science things.
[14:18] GenAI teaching kit, this one might be
[14:21] the one we're most proud of for a couple
[14:22] of reasons. This is the latest
[14:25] this the
[14:26] Before this week this was the latest kit
[14:28] that we had developed. Built this one
[14:30] with a
[14:31] very motivated professor at Dartmouth,
[14:34] Sam Raman, fantastic guy.
[14:36] And
[14:37] one of the unique things about this
[14:39] teaching kit is that not only does it
[14:42] include a number of those self-paced
[14:44] courses that provide the course
[14:46] certificates for students when they
[14:47] complete them.
[14:49] This one also prepares students for a
[14:52] much more involved, much larger
[14:54] certification exam that that the DLI
[14:57] offers as well and I'll I'll talk about
[14:59] the certification exams in a second.
[15:02] But the exam prep is completely aligned
[15:05] with the content in this kit.
[15:07] >> [snorts]
[15:09] >> And and again, so we're announcing here
[15:12] today in this session for the first time
[15:14] the release of the open USD teaching
[15:16] kit.
[15:17] Open USD has to do with 3D graphics and
[15:21] simulation. There's some fundamental
[15:23] aspects of physical AI
[15:26] in this one as well and this one also
[15:28] aligns with the new open USD
[15:31] certification exam as well. So again,
[15:33] you could be teaching students
[15:35] throughout the curriculum using these
[15:37] teaching kits and in parallel be
[15:40] preparing them to take these
[15:41] certification exams which are becoming
[15:43] more and more popular again with with
[15:45] employers and making the students a lot
[15:47] more attractive out in in industry.
[15:51] Okay, so this is the certification
[15:53] exams. So these are exams that are
[15:56] proctored online.
[15:59] This is the only thing that actually has
[16:01] a cost with any of these programs again
[16:04] just because we're doing this through a
[16:05] third party proctoring company called
[16:08] Certiverse if you've heard of them.
[16:10] We have a number of certification exams
[16:13] both at the associate level and the
[16:16] professional level. So with students the
[16:19] associate
[16:20] exams are probably going to be a little
[16:23] more appropriate. But
[16:25] yeah, and we we've been offering these
[16:28] exams for free here at GTC as well.
[16:31] So the Nvidia training organization is
[16:33] in the has been in the Signia Hilton
[16:36] Hotel. We used to be in the conference
[16:37] center
[16:38] but the show has grown so much that
[16:41] we're in the hotel now and today being
[16:44] the last day I
[16:45] would assume that there's still probably
[16:47] some slots open. But yeah, you can just
[16:49] go in and and try out an exam if you
[16:51] want for free. There's really nothing to
[16:52] lose.
[16:53] So if you haven't done that, I encourage
[16:54] you to stop by
[16:55] and check that out.
[16:59] So these are the magic links to these
[17:02] two programs. I encourage you guys to
[17:05] to check out both of them. Again,
[17:07] teaching kits is is the easiest way to
[17:09] get started. It's just a 5-minute
[17:11] application that you guys can fill out
[17:14] and then probably get to applications on
[17:17] Monday
[17:18] approvals for those. I'll try to
[17:19] actually do them tomorrow as well.
[17:22] But we'll see.
[17:23] And then again, the ambassador program
[17:26] that one again is a
[17:28] little more involved application there
[17:30] because we do ask for a really short
[17:34] resume style documentation that shows
[17:36] that you have some of the background
[17:38] relevant to the workshops you want to
[17:40] get certified to teach. So
[17:43] All right. And
[17:45] with that, I am going to turn it over to
[17:48] Professor Paul Nussbaum who has been
[17:51] really a champion in using these
[17:53] programs, sort of a pioneer especially
[17:55] with the new certification exams.
[17:58] But Paul is certified to teach a number
[18:00] of these workshops to his students and
[18:03] he's going to talk a little bit about
[18:05] you know, student feedback and and and
[18:07] how these programs have been a real
[18:09] success in your university. So Thanks so
[18:11] much, Joe. Appreciate it.
[18:13] All right, thank you very much.
[18:15] Well, thanks for coming out. I
[18:16] appreciate that. I did want to talk a
[18:18] little bit about
[18:20] uh
[18:21] how we offer these workshops to our
[18:24] students and [snorts] how incredibly
[18:27] timely it is now to
[18:30] to go ahead especially students that are
[18:32] very you know, focused on their future
[18:33] careers. They're very very nervous. You
[18:36] don't need that. There you go. Thank
[18:37] you.
[18:38] They're very nervous about
[18:40] you know, future career prospects as
[18:43] undergraduate students and graduate
[18:45] students may be.
[18:47] So as a career focused university, ECPI
[18:51] does really take quite a bit of time to
[18:53] make sure that
[18:55] we're preparing our students for a
[18:58] future workforce.
[19:00] Um
[19:02] We have an accelerated program so no
[19:04] football teams but
[19:07] you know, very focused
[19:10] just going ahead and graduating as
[19:12] quickly as you can. We have a very
[19:14] structured program that allows students
[19:17] to get through the curriculum very very
[19:20] quickly in an accelerated way. And we
[19:23] have campuses mostly on the Mid-Atlantic
[19:26] but also in Florida and Texas.
[19:30] And in general
[19:32] if talk about AI training, we feel that
[19:35] it's critical for our students to really
[19:37] be prepared for a workforce where they
[19:40] may be creating these AI solutions or
[19:44] they may just be using those AI
[19:46] solutions.
[19:48] So in terms of our strategy, it's kind
[19:50] of a
[19:50] uh uh multi-faceted
[19:52] strategy as you can see here.
[19:54] Uh we have obviously academic programs
[19:57] in artificial intelligence and machine
[19:59] learning, but also hands-on
[20:02] certifications and workshops that Joe
[20:04] just mentioned.
[20:06] As uh our ongoing collaboration with
[20:08] Nvidia's proved very very fruitful in
[20:11] that regard. So, uh we also have other
[20:14] activities in terms of interdisciplinary
[20:16] learning.
[20:17] Um I'll talk to you a little in a moment
[20:19] about how we work with our culinary
[20:22] students and kind of show them how
[20:24] artificial intelligence can impact their
[20:27] future careers. But, uh we feel that AI
[20:30] is going to impact every uh future
[20:32] career. So, that's kind of why we take a
[20:35] holistic approach.
[20:37] In terms of academic programs, we do
[20:40] have quite a few uh curriculum
[20:44] development is ongoing, obviously, but
[20:47] you know, primarily in the computer
[20:48] science and where I teach electronics
[20:51] engineering, electronics engineering
[20:53] technology.
[20:54] So, um
[20:56] these are really once again, as I
[20:58] mentioned ECPI University being
[21:00] career-focused, a lot of these are AI
[21:03] courses, but also to prepare you
[21:06] uh to work in industry in robotics and
[21:09] manufacturing automation. These are
[21:11] things that
[21:12] uh areas where a great deal of hiring is
[21:15] taking place, and we expect that to
[21:17] grow. We expect industrial automation to
[21:21] only grow in in size as uh as uh we kind
[21:24] of bring more of these facilities
[21:27] onshore, the big onshoring effort uh
[21:30] here in the United States. So, um I do
[21:33] offer the 8-hour workshops as an Nvidia
[21:36] ambassador. That's one of the things
[21:39] that I do, and students have to take
[21:41] time out of their busy personal
[21:44] schedules to go above and beyond the
[21:47] coursework and sit in on one of these
[21:50] 8-hour workshops. And I tell you what,
[21:52] they they sell out every time. I'm only
[21:56] allowed to offer uh the workshop to 20
[21:59] attendees, but for every teaching
[22:02] assistant that I have working with me on
[22:04] that day, I can add another 20
[22:07] attendees, and I regularly uh fill up
[22:10] 100 uh student capacity workshops
[22:13] regularly.
[22:15] They're very attracted to the Nvidia
[22:17] brand, and
[22:19] the same for myself. They really they
[22:22] know the Nvidia brand from their gaming,
[22:25] you know, perhaps or just popularization
[22:27] of the Nvidia accelerator boards for
[22:30] personal use,
[22:32] but they also respect the Nvidia brand
[22:34] for its quality. And they know that if
[22:37] they earn a numbered and traceable
[22:40] certificate from Nvidia, that it is
[22:43] something that carries some weight as
[22:45] part of their portfolio, resume,
[22:48] LinkedIn profile, etc. So, that's why I
[22:52] enjoy being an ambassador delivering
[22:54] these Nvidia content because I know I'm
[22:57] going to get a quality workshop. I know
[23:00] I'm going to get a workshop where you
[23:01] cannot cheat to pass the assessment. I
[23:05] know I'm going to get a workshop that's
[23:06] going to be respected
[23:08] after the student earns that
[23:11] certificate. So, very very popular, and
[23:15] like I said, they're doing it on their
[23:17] spare time. They're they're taking these
[23:19] workshops outside of class hours. So,
[23:22] I'm very very excited about our
[23:24] wonderful ECPI University students, but
[23:27] I think you also can be that hero for
[23:30] them, and that's you know, a bit of a
[23:32] what I'm trying to share with you. It's
[23:34] really been a great benefit to my
[23:37] personal career advancement, but also
[23:40] for our students
[23:41] just being able to be an ambassador, and
[23:43] I would encourage all of you to at least
[23:45] you know, think about going through that
[23:47] process and and becoming an ambassador
[23:50] at your university.
[23:53] So, in terms of workshop, I've I've
[23:55] personally taught over 1,000 students
[23:57] since 2022 going through these
[24:00] workshops. I
[24:02] predominantly offer the fundamentals of
[24:04] deep learning because it is a
[24:05] prerequisite
[24:07] for other more advanced workshops.
[24:10] Students know they have to pass that
[24:12] workshop before they can take one of the
[24:14] other more advanced workshops, and there
[24:17] they learn to build, train, and test a
[24:20] variety of different pattern analysis
[24:23] machine learning
[24:24] generative models. So, they're they're
[24:27] doing hands-on. I I tell students, this
[24:30] is not a lean-back webinar that you
[24:33] attend for 8 hours. This is a
[24:35] lean-forwards
[24:36] hands-on. You've got to modify code,
[24:39] you've got to read code, you've got to
[24:41] train. The nice thing is is you're
[24:44] driving a Ferrari.
[24:46] Nvidia provides the GPUs, and boy is it
[24:50] nice and zippy when you go home and try
[24:52] that on your laptop. It's a little bit
[24:54] slower to train the same models, isn't
[24:56] it?
[24:57] Um
[24:57] we basically uh offer a a number of
[25:00] other ones that have been popular.
[25:02] Anomaly detection is very popular,
[25:05] especially with our cybersecurity
[25:07] students and the onset of zero-day
[25:10] attacks. It is focused on just munching
[25:14] through millions of lines of log files
[25:17] to find unusual or anomalous activity.
[25:21] So, cybersecurity students really do
[25:24] like to attend that workshop. Predictive
[25:27] maintenance is being restructured right
[25:29] now. You can find more details on those
[25:32] websites that
[25:34] Mr. Bongo has pointed you to. It's right
[25:36] now it's being reworked into a a larger
[25:39] scope, but that is very popular with our
[25:42] students interested in industrial
[25:44] automation.
[25:46] The
[25:47] diffusion model is just a great great
[25:49] fun workshop. You know, how do I turn
[25:52] text into a picture? Right? I mean,
[25:55] that's people play with it online, but
[25:57] they want that just that hungry
[26:00] curiosity that electronics engineering
[26:02] students, at least my students, they're
[26:03] just hungry to learn how does that work,
[26:05] right? So, [snorts] they love that one,
[26:07] but I know also those diffusion models
[26:10] are going to be very important for us in
[26:12] terms of motor torque control when we
[26:15] move into the physical AI. So, I I love
[26:18] teaching it because it shows them under
[26:20] the hood how these diffusion models
[26:22] work. And then the Jetson workshops are
[26:25] a little bit lighter. They don't require
[26:27] Python programming, but they do teach
[26:30] you how to train and collect the data to
[26:34] train a brand new neural network or
[26:37] using transfer learning to retrain it.
[26:40] So, here you can see one of those Jetson
[26:42] workshops. They don't have to be
[26:44] electronics engineers. They can just be
[26:46] curious about AI, and you can see here
[26:50] uh this silly picture of me bending
[26:54] forwards. I've got kind of a crazy plaid
[26:56] Madras shirt on because what they're
[26:59] going to be doing is they'll be training
[27:01] a neural network to recognize thumbs up
[27:04] versus thumbs down, as well as a number
[27:06] of other exercises. And so, they'll be
[27:08] collecting that data from different
[27:10] angles, right? [snorts] Different
[27:12] positions, and then I'll walk over and
[27:15] say, "Great, show me that workshop." But
[27:17] of course, I'm wearing that crazy
[27:18] pattern shirt that they did not use in
[27:21] training, and we'll learn these
[27:23] fundamental aspects about the importance
[27:25] of the variation of the training data.
[27:29] So,
[27:30] I really love the ambassador program,
[27:33] and our students love it. They're making
[27:35] wonderful testimonials,
[27:37] you know, telling me about in the early
[27:39] days how much they enjoyed taking the
[27:42] workshops, and then more importantly,
[27:45] more recent testimonials, how it's
[27:47] really impacted their careers. How
[27:50] they've been able to get into the
[27:52] master's degree program, how they've
[27:53] been promoted to lead AI development
[27:56] teams at their organizations because
[27:59] they have the numbered and traceable
[28:01] Nvidia certificate that they've earned,
[28:03] and nobody else on the team had it,
[28:05] right? So, they they really love that.
[28:08] So,
[28:10] we go also above and beyond the
[28:13] the certifications do require a sequence
[28:17] of workshops. So, we're going to be
[28:19] introducing that coming forwards. We
[28:21] have a We just launched a certification
[28:24] portal for our students.
[28:27] Allows them to more easily understand
[28:29] what the prerequisites are, how to
[28:30] prepare for certifications.
[28:33] And we're studying the Nvidia teaching
[28:35] kits. As I said, ours is an accelerated
[28:38] program, so sometimes we just need to
[28:40] take slices from that
[28:43] teaching kits. But as I mentioned, these
[28:46] sort of things come together. We take
[28:47] our students who have passed these
[28:49] workshops, we pair them together with
[28:51] students who know nothing about
[28:53] artificial intelligence, and we talk
[28:55] about things like the Chef's Buddy that
[28:58] creates recipes and draws plating images
[29:01] running continuously in our electronics
[29:04] laboratory. So, it it it lets the
[29:06] students kind of see how they might help
[29:09] someone else with their AI training. So,
[29:12] don't forget interdisciplinary. I know
[29:15] we sometimes have a kind of a siloed
[29:17] version. We also offer AI competitions
[29:21] and industry collaboration as many
[29:23] universities do. You're probably already
[29:26] familiar with that, but just in general,
[29:30] all of our curriculum at ECPI University
[29:32] is being weaved with AI, not just the
[29:36] ambassador program. We're also using AI
[29:39] even in our business degree. We We need
[29:42] our students to play with it, have
[29:43] hands-on. ECPI is a career-focused
[29:46] university. We need our students to be
[29:48] ready to hit the ground running when
[29:50] they start in those careers. I
[29:52] personally will be focusing on agentic
[29:55] AI and probably all of us will, won't
[29:57] we? And
[29:59] so I'll be adjusted as credentialed to
[30:02] teach two more workshops building rag
[30:05] agents using LLM and AI agent
[30:08] development using LLM. Great great
[30:10] workshops looking forward to teaching
[30:13] them. But listen, I really thank you for
[30:15] your time. It's wonderful that you're
[30:16] here with us today and hopefully I've
[30:18] gotten you interested and pumped. Just
[30:20] like you're excited to be here at the
[30:22] show, you can get your students excited
[30:25] to learn about these material, okay?
[30:29] All right, thank you SO MUCH.
[30:31] >> [applause]
[30:35] >> WE'RE GOING TO SAVE ALL THE Q&A.
[30:36] >> OH, QUESTIONS. Very good, yeah.
[30:38] Huh?
[30:39] One more slide, I'm sorry.
[30:41] Yes, those websites very important.
[30:44] >> [sighs and gasps]
[30:45] >> All right, thank you. Thank you Paul.
[30:47] Thank you Joe. Now we have time for a
[30:49] couple questions. I can come around and
[30:52] give the mic to anyone that has
[30:53] questions for Paul and Joe.
[31:01] All right, I'm going to go right ahead.
[31:02] Next.
[31:09] Hi, thank you very much. Um this is a
[31:11] question for Paul. Um regarding those
[31:14] eight-hour workshops that you have sort
[31:16] of included as part of your curricula.
[31:19] So we're not part of the coursework.
[31:23] Okay. They're separate, yes. Okay, so so
[31:25] my question was really
[31:27] do they have to do the assessment sort
[31:29] of in that same eight hours or do they
[31:31] allow to do it sort of whenever Nvidia
[31:34] allows them to do so?
[31:35] He'll explain it better. So so the you
[31:37] can break the six to eight-hour
[31:39] workshops up into multiple days. So most
[31:42] most a lot there
[31:44] teaching those in the curriculum in the
[31:47] class meeting time is a very popular use
[31:49] case.
[31:50] And so in that case you could just split
[31:52] the workshop into say one and a
[31:54] half-hour chunks or one hour chunks.
[31:56] We're really flexible in allowing that.
[31:59] You just have to make sure that the
[32:00] students stop their GPU instances
[32:03] before before moving on to the next
[32:05] class cuz sometimes they'll they'll
[32:06] continue to run on the back end and burn
[32:08] up a lot of GPU time. But but yeah, most
[32:11] assuredly yes, we're very flexible in
[32:12] that aspect.
[32:16] Okay, good morning. Thank you for
[32:17] attention. I'm Jackie from Hong Kong
[32:19] Polytechnic University. So we are very
[32:22] interested to introduce this program to
[32:24] our university but
[32:26] I want to know more is because is we
[32:28] find difficulties how to integrate or
[32:31] embed this in the academic subjects.
[32:34] The teaching kits are a great way to do
[32:37] that because they're already developed
[32:39] and you can pick and choose which
[32:40] modules you want to incorporate into
[32:42] your curriculum for a particular course.
[32:45] Yes. Um that's pretty much
[32:47] >> Yeah, for for the most part we leave it
[32:49] up to you guys to decide how you want to
[32:51] do it. So we provide again just lots and
[32:54] lots of material in a a bunch of
[32:55] different formats and for the most part
[32:57] let you guys decide how you want to
[32:59] integrate it into your class.
[33:00] >> Maybe can I share is in a case you offer
[33:03] this as a co-curricular activities for I
[33:05] do. I say if you're willing to attend on
[33:09] a Friday
[33:10] where we we normally our classes are
[33:12] Mondays to Thursdays. We say earn a
[33:15] numbered and traceable certificate from
[33:18] Nvidia in eight hours. These people are
[33:20] very busy. They've got families, they
[33:23] love that.
[33:25] Do they all earn it? I've got about a
[33:27] 70% pass rate on many of these workshops
[33:30] but they all learn and they all love it
[33:32] even if they don't necessarily want to
[33:34] become a Python programmer. Yes, I want
[33:37] to learn more about now if you want to
[33:39] in start doing this maybe can share
[33:42] because you gone through process what
[33:43] will be the major challenge the first
[33:45] one. And the second one for this would
[33:48] you charge students or I means the this
[33:50] is additional resources is paid by the
[33:53] department or somewhere else, yeah.
[33:55] Thank you. Um I just my own time. I
[33:59] I don't know there's no fee to the
[34:01] students. Yeah, the It's your personal
[34:05] contribution to this training. Yes, yes.
[34:08] I it's you know as a
[34:10] it's a profession, you know, it's part
[34:11] of my advancement of my profession so I
[34:14] uh
[34:15] If we want to scale up, we'd like to
[34:17] mobilize mobilize our faculty members or
[34:20] other instructors to join it. So that's
[34:21] why we also have to work experience in
[34:24] the matter. Yeah, sometimes
[34:26] sometimes we give you know gifts or gift
[34:28] cards to the teaching assistants but not
[34:31] not me, no. And then we have to
[34:33] talk them more for sure. Yeah, I can
[34:35] answer more questions.
[34:36] Oh, thanks.
[34:39] Um
[34:40] I had sort of a two-part question. The
[34:42] first is about the ambassador program.
[34:44] I'm a staff member at UT Austin and I
[34:47] saw in the slide about the ambassador
[34:49] program it specified faculty at
[34:51] universities. Is there the capability to
[34:54] become an ambassador as a staff member?
[34:57] It is especially if you are as part of
[35:00] your job training and teaching and
[35:03] running workshops.
[35:05] If that's the case then then yeah.
[35:07] The real the real requirement is a level
[35:09] of employment. So again then from there
[35:12] we would sort of look at what you're
[35:13] what you're doing on a daily basis and
[35:15] do we do we think that you're going to
[35:17] be teaching and have the opportunity
[35:20] to teach teach workshops and have access
[35:22] to the audiences and and all of that.
[35:24] Okay, awesome. Thank you.
[35:26] And I forgot the second question so I
[35:27] think that's it.
[35:32] Last one and then we really do have to
[35:33] close it. Meet him in the back there.
[35:36] Hello.
[35:38] I'm coming from Dominican Republic
[35:40] and
[35:41] I am the CEO of of an innovation lab. We
[35:43] working in AI and robotics and
[35:48] I assessing two of the main
[35:50] universities, the biggest university
[35:53] elite university of United States and
[35:55] sorry, Dominican Republic. So the idea I
[35:58] have built with them a minor.
[36:01] We minor in AI and emerging technologies
[36:04] applied to every career. So the lawyer
[36:08] the lawyer will be out with an minor
[36:11] with AI applied to the to the lawyer the
[36:13] professional and the doctor and
[36:15] everyone. This is one more semester at
[36:18] the end of all the program. We are ready
[36:21] to launch even the Saturday we will be
[36:24] launching that program in the in the
[36:26] Dominican Republic. So I would like to
[36:28] see if is any way to collaborate because
[36:31] you have
[36:33] a road
[36:34] down and maybe Nvidia we also we I I I I
[36:38] have contact with the rect with the the
[36:42] principal of the of the both
[36:43] universities so it will be very easy to
[36:45] me to incorporate
[36:47] all of all of these to to that program.