People · Co-founder & CTO · Simulation & learning

Matt Bennice.

Simulation-driven robot control, imitation and reinforcement learning at fleet scale, and the sim-to-real discipline behind robots that work outside the lab. Google X, Everyday Robots, DeepMind, Wing, Stack AV, 1X, Cruise.

Simulation-driven controlImitation learningReinforcement learning at scaleSim-to-real transferFleet learningHuman-robot interaction

LinkedIn ↗·dblp·US11938638B2

01 / Selected publications

The papers.

Learning to Learn Faster from Human Feedback with Language Model Predictive Control
Robotics: Science and Systems · 2024
Deep RL at Scale: Sorting Waste in Office Buildings with a Fleet of Mobile Manipulators
Robotics: Science and Systems · 2023
Jump-Start Reinforcement Learning
ICML · 2023
Practical Visual Deep Imitation Learning via Task-Level Domain Consistency
ICRA · 2023
VADER: Visual Affordance Detection and Error Recovery for Multi Robot Human Collaboration
arXiv · 2024
Interactive Multi-Robot Flocking with Gesture Responsiveness and Musical Accompaniment
ACM Transactions on Human-Robot Interaction · 2026

Full record on dblp ↗

At Cereal

At Cereal, Matt owns the world loop: the always-on simulation that mirrors the home, verifies every plan before a body moves, and turns recorded operation back into training and verification data.

One mind. Any body.