Back to full agenda
Track 2: Push Models Further
Hands-on lab
Let's Get Physical with AI: Teaching Robots New Skills in Simulation
Session
LAB04
Day & Date
Time
Location
About this session
Experience the complete workflow for adapting a robotic action policy with reinforcement learning. Configure a training strategy, run it on CoreWeave infrastructure, test the resulting policy in simulation, and use Weights & Biases to understand and improve its performance. It's structured as a competition — teams closest to the success criteria win.
Speakers
Anu Vatsa
Share this session
Get in the room. San Francisco, September 29.
Fully Connected 2026 is where the engineers, leaders, and operators running AI in production come together for three days of depth, access, and real conversation.