Motoniq · 1 day ago
Head of AI (VLA)
United States
Full-time
Remote
Director/ExecutiveMotoniq is building the full-stack data engine for Physical AI, delivering high-fidelity, robot-ready human motion data at scale. They are seeking a fractional Head of AI to architect their AI strategy and ensure data products integrate with embodied AI systems, requiring deep expertise in VLA and robotic learning.
Computer Software
Responsibilities
Define technical requirements and data specifications to ensure compatibility with major VLA architectures (RT-2, Octo, OpenVLA, π₀, and emerging models)
Establish validation frameworks to verify data quality meets training requirements
Track developments in embodied AI, foundation models for robotics, and multimodal learning
Advise on which research directions will most impact our data product strategy and customer needs
Identify opportunities to contribute to open-source VLA ecosystems
Design and implement validation systems for sensorimotor data streams
Develop metrics for data quality, diversity, and distribution alignment
Ensure annotation, augmentation, and packaging processes preserve information critical for policy learning
Engage directly with robotics companies and research labs to understand their VLA training workflows, data requirements, and integration challenges
Translate technical feedback into product improvements
Represent Motoniq at conferences, publish technical documentation, and build relationships within the embodied AI research community
Establish credibility as the authoritative source on training data for Physical AI
Qualification
Required
PhD in Computer Science, Robotics, Machine Learning, or related field with focus on embodied AI, robotic learning, or vision-language-action models
Hands-on experience training or deploying Vision-Language-Action (VLA) models such as RT-1, RT-2, Octo, OpenVLA, PaLM-E, π₀, or similar architectures
Deep understanding of robotic learning paradigms: imitation learning, behavior cloning, reinforcement learning, and foundation model fine-tuning for manipulation tasks
Experience with multimodal data (vision, language, proprioception, tactile) and sensor fusion for robotics applications
Track record of publications in top-tier conferences (CoRL, RSS, ICRA, IROS, NeurIPS, ICLR, CVPR) or contributions to major open-source robotics/AI projects
Strong software engineering skills in Python, PyTorch/JAX, and robotics stacks (ROS, Isaac Sim, MuJoCo, or similar)
Based in San Francisco Bay Area or willing to relocate post-funding
Preferred
Previous work at leading robotics/AI labs: Google DeepMind, Google Research (Robotics), OpenAI, Tesla AI, Figure AI, Physical Intelligence, UC Berkeley RAIL Lab, Stanford SVL, CMU Robotics Institute, MIT CSAIL
Experience building or contributing to large-scale robotic datasets (e.g., Open X-Embodiment, RoboNet, Bridge Data, DROID)
Expertise in human motion capture, biomechanics, or human-robot interaction systems
Understanding of data markets, ML data pipelines, or experience at data-centric AI companies (Scale AI, Robust Intelligence, Snorkel AI)
Active contributor to embodied AI open-source ecosystems (OpenVLA, Dora-rs, Open Robotics, LeRobot)
Benefits
Competitive equity package aligned with senior technical leadership contribution.
Relocation support provided for Bay Area move post-funding.
Company
Motoniq
Motoniq is an AI Robotics Company building motion intelligence for humanoid robots - enabling machines to learn directly from human movements Through its novel Motion-Capture platform, Motoniq is the first to record real human motion and tactile data to train AI systems with human dexterity, precision and adaptability.
Founded in 2025
2-10 employees
Funding
Current Stage
Early StageCompany data provided by crunchbase