Junior Research Scientist

The AI Robot Association (AIRoA) is collecting one million hours of humanoid robot operation data with hundreds of robots to train Vision-Language-Action (VLA) models, and this junior research scientist position joins that effort. It works on the same tasks as the senior research scientist role within an open, shared robot data ecosystem.

The scientist builds data preprocessing pipelines for multimodal robot datasets and trains VLA models using supervised learning, reinforcement learning, fine-tuning and RLHF. Models are developed and evaluated both in simulation and on physical robots and deployed to humanoid and mobile platforms. It also involves improving training efficiency and publishing at leading conferences.

Required qualifications are an MS with three or more years of industry experience, or a PhD, in computer science, electrical engineering or a related field; a first-author publication at a top venue such as CoRL, ICML, CVPR or NeurIPS; experience with reinforcement learning and training LLMs, VLMs or VLAs; work with simulators or real robots; and deep learning frameworks such as PyTorch. Preferred are robot learning research, tactile sensing, MuJoCo or Isaac Sim, and ROS.