About the Training Lab
Every robotics team faces the same challenges: data scarcity and tedious development processes. This manual, tool-by-tool workflow is the reason moving from sim-to-real is still mostly for specialists with months to spare, even when great algorithms and the necessary compute is available.
Join this workshop to learn a repeatable methodology, cocreated with a special agent that understands a specific set of skills to operate the entire pipeline. Rather than learning five separate simulation and ML tools, participants interact with a single conversational interface that orchestrates the full stack.
What You'll Build
Each participant will have access to a dedicated HP ZGX Nano AI Station—a personal AI supercomputer with 128 GB unified memory and 1 PFLOP of AI compute. The instructor will guide attendees through a robotic automation workflow for a hospital to see how to set up a simulation environment, compose a digital twin of the hospital, and capture expert workflow data to generate synthetic variations. Then participants will train or fine-tune a physical AI policy, and validate the workflow with an agent-first approach before real-world deployment.
Takeaways include:
- An agent-ready hospital automation workflow
- A digital hospital scene with context
- A task-oriented simulation environment (Isaac Sim™, Isaac™ Lab-Arena)
- Synthetic data variations for edge cases
- A trained/fine-tuned policy (GR00T, Isaac Lab, reinforcement learning)
- A validation view showing digital agents observing, reasoning, and triggering actions
Prior Experience Helpful
Participants should have familiarity with OpenUSD, NVIDIA Omniverse™, simulation concepts, healthcare AI data workflows such as MONAI and MAISI, and physical intelligence models like GR00T, as well as an OpenAI or Claude Code API key.
Prior Isaac Sim or Isaac Lab experience is helpful.
Registration
Two time slots offered:
- Wednesday, June 24, 10:00–14:00 Lunch break (13:00–13:30)
- Thursday, June 25, 10:30–14:30 Lunch break (12:45–13:15)
Space is limited; early registration is recommended.
This event is exclusively open to confirmed Hamlyn Symposium ticket holders.
The training lab registration deadline is June 15th, and we will send out attendance confirmations via email by June 20th.
Following approval, look for a confirmation email containing additional training lab details. If your plans change, please notify us so we may offer your spot to another guest.
Agenda
Introductions and check access, environment assumptions, and the agent-first goals for Isaac for Healthcare
See how the Rheo blueprint within Isaac for Healthcare leverages agent-ready, skill-based development for hospital automation scenarios. Understand how Project Rheo turns hospital spaces and workflows into sim-ready, trainable digital environments.
Set up the simulation environment, import robot and rigging, perform data collection/multiplication/augmentation, implement policy training.
Speakers
Maximilian Ofir
Technical Marketing Engineer
NVIDIA
Mostafa Toloui
Product Management Lead
NVIDIA
Mikael Budfors
Solution Architect
NVIDIA
Wednesday, June 24, 10:00 am or Thursday, June 25, 10:30 am
NVIDIA Training Lab at Hamlyn Symposium
FAQ
Who is this training lab for?
Who is this training lab for?
This hands-on lab is for robotics developers, computer vision ISVs, simulation engineers, and technical teams building healthcare robotics or hospital automation systems and workflows. Registration for the Hamlyn Symposium is required in order to sign up for this training lab.
What will attendees learn?
What will attendees learn?
Attendees will learn how Isaac for Healthcare 0.6 and Project Rheo support an agent-ready workflow: set up simulation, compose a digital hospital environment, generate data, train a model or policy, and validate before deployment.
Is this a product overview or hands-on lab?
Is this a product overview or hands-on lab?
It is a hands-on technical lab with product context.