About the Company
At VinDynamics, we design safe, affordable, and intelligent humanoid robots to assist in everyday life — robots for everyone. Backed by Vingroup, Vietnam's leading technology conglomerate, we are on a mission to make advanced robotics accessible, reliable, and beneficial for billions of people worldwide. By combining cutting-edge AI, world-class engineering, and human-centered design, we aim to seamlessly integrate robots into daily life — enhancing safety, productivity, and happiness at home and beyond.
Key Responsibilities
- Develop and implement reinforcement learning algorithms specialized for locomotion tasks (e.g., walking, running, climbing, balancing).
- Design, integrate, and optimize high-fidelity simulation environments for safe and efficient policy training.
- Conduct sim-to-real transfer by addressing robustness, domain randomization, and system identification challenges.
- Incorporate perception, sensor feedback, and proprioception into RL agents to enable adaptive and reactive motion.
- Evaluate and benchmark locomotion policies under diverse real-world conditions (e.g., terrain variation, disturbances, slopes, payloads, and friction).
- Work on reward design, stability, sample efficiency, and safety-constrained learning.
- Write clean, maintainable, and well-documented code, ensuring reproducibility and version control for experiments and policies.
Requirements
- Solid background in Reinforcement Learning (Deep RL, Policy Gradient, Model-based RL, Imitation Learning, etc.).
- Hands-on experience with simulation platforms such as MuJoCo, PyBullet, Isaac Gym, or Gazebo.
Preferred Qualifications
- Experience with locomotion, motion control, or physical control systems (e.g., legged robots, drones, exoskeletons, robotic arms).
- Experience in sim-to-real transfer, domain randomization, or system identification in robotics.
- Proficiency in Python and/or C++, and familiarity with ML frameworks such as PyTorch, TensorFlow, or JAX.
- Strong analytical and debugging skills for physical systems; ability to identify stability and performance bottlenecks.
- Familiarity with sensor fusion, feedback control, and proprioceptive sensing.