Your Responsibilities

• Design and train foundation models that integrate vision, language, and actions for embodied intelligence
• Adapt LLMs and VLMs for robotic control, planning, and interactive behavior, enabling context-aware decision making
• Develop AI-driven control policies for manipulation, grasping, and motion planning using reinforcement learning, imitation learning, and foundation model approaches
• Build modular, scalable, and high-performance data processing, training, and inference pipelines for large-scale datasets
• Design reproducible workflows for training, evaluation, and deployment, including benchmarking for generalization, safety, and task success
• Push the state of the art and bring it into production — this includes staying current with (and ideally contributing to) the research literature, and translating advances into systems that actually ship
• 
  

Technologies

• Python, C++
• PyTorch (or TensorFlow / JAX)
• Transformer-based models (LLMs, VLMs, multimodal architectures)
• Distributed training frameworks
• Linux, Docker (Kubernetes nice to have)
• ML pipelines and model deployment systems
•