Louis Le Lay

I'm Louis Le Lay. I build learning-based controllers that make humanoid robots walk and move.

At UMA, I own locomotion and simulation for ASTRO, taking Version 0 from bare hardware to walking and whole-body teleoperation. I also deploy the same stack on platforms including the G1 and X2.

Previously, I built the first RL controller to walk PAL Robotics' Kangaroo humanoid, replacing its legacy MPC walker. I am also a co-author of mjlab and an active contributor to Isaac Lab.

GitHub · LinkedIn · Twitter · E-mail · Scholar

ASTRO Version 0 walking during its reveal by UMA Robotics — ASTRO Version 0 reveal — locomotion & whole-body teleoperation

Kangaroo humanoid walking on real hardware with RL locomotion — Kangaroo at 0.5 m/s — RL replacing 20 years of MPC

G1 robot getting up using reinforcement learning — G1 fall recovery

Tennis motion retargeting to a robot — Motion retargeting

Experience

UMA · Research Engineer Mar 2026 – Present

Own locomotion and simulation for ASTRO — bare hardware to walking and whole-body control. This work was featured in Rémi Cadene's keynote at MACHINA Summit 2026. Also develop locomotion and whole-body control for the G1 and X2.

PAL Robotics · Robotics Engineer Jul 2025 – Mar 2026

Built the full RL sim-to-real locomotion pipeline for the Kangaroo humanoid, replacing the legacy MPC-based walker.

University of Tokyo · Research Intern Jun – Sep 2024

Airbus · Robotics Engineer Intern Sep 2023 – Jun 2024

Publications

For a complete and up-to-date list, see my Google Scholar.

mjlab: A Lightweight Framework for GPU-Accelerated Robot Learning

Kevin Zakka, Qiayuan Liao, Brent Yi, Louis Le Lay, Koushil Sreenath, Pieter Abbeel

arXiv preprint, Jan 2026

paper · website · code

Software

pal_mjlab

PAL Robotics' robots in mjlab. Standard RL interface for Kangaroo, Talos, and TIAGo.

code ·

mjlab

Isaac Lab API on MuJoCo Warp. Led by Kevin Zakka.

code ·

Kinova Gen3 RL & Sim2Real Toolkit

Isaac Lab training to real Kinova Gen3 deployment via ROS 2.

code ·

SO-ARM RL: Manipulation tasks

RL manipulation tasks for SO‑ARM100/101 in Isaac Lab.

code ·

Press & Mentions

MACHINA Summit 2026 — Rémi Cadene Keynote — ASTRO Version 0's locomotion and whole-body teleoperation were featured in the keynote. Jul 2026. [ video ]
UMA / ASTRO Version 0 — Selected Coverage — Coverage of UMA's first humanoid prototype, including its locomotion, push recovery, and whole-body teleoperation. Jul–Aug 2026. [ BFM Tech · Journal du Net · French Tech Journal · L'Embarqué ]
BFM Business — Les Experts de l'IA — Olivier Stasse (Research Director, LAAS-CNRS) cited me as an example of rapid progress in humanoid robotics. Jun 2026. [ video ]
VivaTech 2026 — Live Demo with PAL Robotics × LAAS — Referenced by Olivier Stasse (Research Director, LAAS-CNRS) for my work on Kangaroo's RL locomotion. Jun 2026. [ video ]

Talks

From Arms to Humanoids: an RL Journey — Invited talk at INIT Robots Lab, Apr 2026.
Teaching a Full-Size Humanoid to Walk with RL, from Scratch — Master's thesis defense, Sorbonne University, Apr 2026.
Training and Deploying RL Agents for Manipulation on the SO-ARM — Talk at ROSCon España, Nov 2025.

Education

Sorbonne University · Master of Engineering, Robotics — 2022 – 2026
ÉTS Montréal · Master of Engineering, Information Technology — 2024 – 2026