Tech Lead, Robotic AI Model
Core
Transform pretrained foundation models into deployable robot policies for manipulation, navigation, and complex physical tasks.
Role type
Tech Lead, Robotic AI Model
Builds
Deployable robot policies for embodied AI systems
Domain
Robotics, Embodied AI, Foundation Models
Deliverable
production ML models
Required skills
Robot learning (imitation, behavior cloning), Post-training paradigms (SFT, RL, diffusion/flow-matching), PyTorch, Distributed training (FSDP, DeepSpeed), Robot simulation (Isaac Sim, MuJoCo, PyBullet, SAPIEN), Sim-to-real transfer, Action representations (continuous control, tokenization, chunking), Python engineering, ROS/ROS2, Real-time control systems
Preferred skills
VLA model fine-tuning (π₀, OpenVLA, RT-2, Octo), Real robot platforms (humanoids, bi-manual arms, mobile manipulators), Large-scale teleoperation data collection, RLHF/DPO/GRPO, Robot learning infrastructure (LeRobot, robomimic, openpi), On-device optimization (TensorRT, ONNX Runtime)
Technologies
PyTorch, Isaac Sim, MuJoCo, SAPIEN, PyBullet, ROS/ROS2, TensorRT, ONNX Runtime
Responsibilities
Design and execute post-training pipelines for VLA and visuomotor policy models; Build and manage large-scale robot demonstration data pipelines; Build and maintain simulation environments for scalable policy training; Build evaluation and benchmarking infrastructure; Optimize models for on-robot inference; Track and adopt state-of-the-art research in robot foundation models
Seniority
Senior, hands-on IC
