Responsibilities:
Design, implement, and train state-of-the-art neural network architectures for robotic applications, including large multimodal models (vision, language, sensor data) and generative models.
Architect and maintain scalable deep learning frameworks for training on massive, real-world robot datasets.
Develop and implement cutting-edge techniques in areas such as imitation learning, reinforcement learning, self-supervised learning, and video-based models.
Own the full machine learning lifecycle: define data collection and labeling strategies, build high-throughput data pipelines, and establish robust evaluation and deployment protocols.
Collaborate closely with software, hardware, and controls teams to integrate and test your models on physical robots.
Leverage and contribute to internal tooling for large-scale data processing, model experimentation, and continuous integration.

Proven experience (M.S., Ph.D., or equivalent industry experience) in Computer Science, AI, or a related field. Strong software engineering skills are non-negotiable; demonstrated excellence in solving complex engineering problems, particularly in Python. A deep, "under the hood" understanding of modern deep learning, including network architectures, loss functions, optimization, and transformers. Extensive hands-on experience with PyTorch. A proven track record of training large-scale deep learning models on large datasets. Ability to thrive in a fast-paced, high-impact environment where solutions are often unclear and require deep exploration. Advantages: Experience with training large models (e.g., LLM, VLM, VLA). Prior experience working with robotic learning systems. Experience with multi-task learning, imitation learning, or reinforcement learning. A strong publication record in top-tier AI or robotics conferences (e.g., NeurIPS, ICML, CVPR, RSS, CoRL).