Rutgers AI/High Performance Computing Workshop

Rutgers AI/HPC Workshop: AI-Driven Research at Rutgers

REGISTRATION IS REQUIRED. SPACE IS LIMITED!

Register to attend: https://shprutgers.jotform.com/262576127488166

Agenda

Part I: The Vision & The Platform

Setting the stage for the new era of scientific discovery.

09:00 AM – 09:30 AM | Welcome & Opening Remarks (Matt Kanoff – Google Antonina Mitrofanova & Leslie Lenert – Rutgers)

• Introductions
• Objectives for the day: From consumption to creation

09:30 AM – 10:15 AM | Session 1: AI at Rutgers (Andy Page – Rutgers OIT & Terrell Moore – Google)

• Google at Rutgers – A peak at existing projects
• AI at Rutgers – Voice of the customer

10:15 AM – 11:00 AM | Session 2: Google Cloud for Science & Gemini’s Science Vision (Adnan Khaleel & Bhanchand Prasad – Google)

• Shift from “AI as a Tool” to “AI as a Partner.”
• The “AI Scientist” concept: Moving beyond chatbots to research agents.
• Case studies: Accelerating literature reviews, hypothesis generation, and experimental design.
• Discussion: How Generative AI is reshaping the scientific method itself.
• Introduction to the Gemini model family

Part II: Demonstrations & Application

Seeing the tools in action before taking the wheel.

11:15 AM – 12:00 PM | Session 3: Demo – AlphaEvolve for HPC Applications – (Adnan Khaleel – Google)

• Live Demo: Evolutionary algorithms optimizing High-Performance Computing (HPC) code.
• Showcase: Automating code discovery and performance tuning.
• Q&A: Integrating Alpha-based tools into existing TAMU workflows.

12:00 PM – 01:00 PM | Networking Lunch (provided)

Part III: Hands-On Labs (The “Builder” Phase)

Interactive sessions where attendees build real AI workflows.

01:00 PM – 01:45 PM | Capricorn demo and discussion (Terell Moore – Google & Dr. Evan Johnson – Rutgers)

• Capricorn: AI-powered framework that utilizes Google’s Gemini models to assist clinicians with systematic literature reviews by cross-referencing patient-specific medical cases with public databases like PubMed in minutes instead of days.

01:45 PM – 02:30 PM | GCP AI Infrastructure, DWS (Bhanchand Prasad – Google)

• Introduction to AI and HPC using GCP

02:30 PM – 02:45 PM | Afternoon Break

02:45 PM – 04:15 PM | Session 4: Lab – Molecular Dynamics and Real-Time Visualization (Adnan Khaleel – Google)

• Objective: How to run NAMD, a molecular dynamics simulation program, on GPUs using the Google Cloud Cluster Toolkit, running the NVIDIA NAMD Container on Slurm.
• Activity: This guide provides instructions on how to run NAMD, a molecular dynamics simulation program, on GPUs using the Google Cloud Cluster Toolkit, running the NVIDIA NAMD Container on Slurm using Apptainer, and using VMD to visualize the molecular structures. Cluster Toolkit is an open-source software offered by Google Cloud
• Outcome: VMD is a molecular visualization program for displaying, animating, and analyzing large biomolecular systems using 3-D graphics and built-in scripting, and using the techniques in this example, users will be able to run their own molecular dynamics simulations, and visualize the 3-D results.

Part IV: Conclusion

04:15 PM – 04:30 PM | Closing & Next Steps (Adnan Khaleel & Matt Kanoff – Google)

• Recap of tools used (Vertex AI, Vizier, Gemini).
• Final Q&A.