- Company Name
- Shaw Daniels Solutions
- Job Title
- Head of Vision – Language – Action (VLA) Development – Manipulation
- Job Description
-
**Job Title**
Head of Vision – Language – Action (VLA) Development – Manipulation
**Role Summary**
Senior technical leader responsible for advancing multi‑modal AI capabilities (vision, language, audio, proprioception, LiDAR, point cloud) in manipulation systems for humanoid robots. Leads strategy and execution of representation learning, behaviour cloning, reinforcement learning, and large‑scale post‑training for VLA/LLM/VLM architectures. Drives end‑to‑end ML pipelines, from data collection to real‑time edge inference, and mentors a high‑performance research and engineering team.
**Expectations**
- Deliver production‑ready VLA models that run on humanoid or legged robot platforms.
- Scale distributed training and optimize models for real‑time inference.
- Build continuous data‑flywheel pipelines for simulation and tele‑operation data.
- Guide the team in research, experimentation, and operational deployment.
- Foster cross‑functional collaboration with MLOps, platform, and robotics teams.
**Key Responsibilities**
- Define and execute strategy for representation learning, behaviour cloning, and RL in manipulation.
- Lead large‑scale post‑training for multi‑modal LLM/VLM/VLA systems.
- Build pipelines to continuously collect, version, label, and evaluate simulation and tele‑operation data.
- Partner with MLOps and platform teams to scale distributed training and optimize models for real‑time edge inference.
- Grow, mentor, and unblock a high‑performing team of researchers and engineers.
- Own end‑to‑end delivery from research to hardware deployment, iterating through data‑driven feedback loops.
**Required Skills**
- 6+ years of deep learning system development, 2+ years in technical leadership.
- Hands‑on design and large‑scale training of LLM/VLM architectures.
- Proven expertise in robotics or autonomous systems: behaviour cloning, actor‑critic, offline RL.
- Experience deploying models to humanoid or legged robots and shipping to real hardware.
- Background in autonomous control and planning.
- Contributions to research or open‑source in multi‑modal transformers, diffusion control, or world models.
- Familiarity with VLA frameworks (e.g., OpenVLA, π models).
- Strong MLOps mindset: distributed training, model versioning, CI/CD for ML.
- Excellent communication, mentorship, and cross‑team collaboration skills.
**Required Education & Certifications**
- Bachelor’s or Master’s degree in Computer Science, Robotics, Electrical Engineering, or related field (or equivalent professional experience).
- Relevant certifications in machine learning, deep learning, or robotics (optional but preferred).