Lead, Vision-Language-Action (VLA) / Behaviour Learning

Kensington Additive

  • London
  • Permanent
  • Full-time
  • 2 days ago
Job Title: Lead Vision Language ActionLocation: London, UKSalary: $100KThis company is pioneering the development of the world’s most advanced, reliable, commercially scalable, and safe humanoid robots. Their first humanoid platform, HMND 01, is a next-generation labour automation unit designed to deliver highly efficient services across multiple domains, starting with industrial applications.What You’ll DoDefine and execute the strategy for representation learning, behaviour cloning, and reinforcement learning (RL).Lead large-scale post-training of multi-modal LLM / VLM / VLA systems, integrating new sensor modalities such as vision, audio, proprioception, LiDAR, and point clouds.Build continuous pipelines that gather simulation and tele-operation logs, maintain a versioned data lake, transform and label data streams with weak supervision, curate balanced datasets, and run evaluation loops that feed new failure cases back into training.Collaborate with MLOps and Data Platform teams to scale distributed training and optimise models for real-time edge deployment.Recruit, mentor, and support a small, high-performing team of research scientists and engineers.We’re Looking For6+ years of experience building deep-learning systems, with 2+ years in technical leadership roles.Hands-on expertise with LLM / VLM architecture design, billion-parameter training, and fine-tuning.Strong background in robotics, autonomous driving, or large language models, with applied experience in behaviour cloning, actor–critic methods, or offline RL.Nice to HaveDeployment experience on humanoid or legged robots.Proven record of shipping to real robots or vehicles and iterating via data-flywheel loops.Experience in autonomous vehicle control and planning.Research contributions or open-source work in multi-modal transformers, diffusion control, or world models.Familiarity with open-source VLA frameworks such as OpenVLA or Physical Intelligence (π) models.What’s on OfferCompetitive salary with participation in a Stock Option Plan.Paid vacation aligned with local labour laws.Travel opportunities to international offices.Office perks: free breakfasts, lunches, snacks, and regular team events.Autonomy to shape the product and lead key initiatives.Collaboration with world-class engineers, researchers, and product experts in AI and robotics.A startup culture built on speed, transparency, and minimal bureaucracy.

Kensington Additive