Machine Learning Engineer (Physical AI)

Location: Bermondsey, London, United Kingdom — fully onsite

Salary: £80,000–£160,000 base salary plus equity

Company

Venture-backed Physical AI startup building models that enable robots to learn and act in the real world.

What you will do

• Fine-tune VLA and world-model checkpoints, progressing toward training models from scratch.
• Accelerate training with efficient dataloaders, sharding, parameter-efficient methods, and experiment sweeps.
• Build evaluation harnesses covering held-out metrics, simulation rollouts, and policy regression testing.
• Develop robot data pipelines, augmentation, and data-quality checks.
• Deploy models onto physical robots and use hardware results to guide subsequent experiments.

Requirements

• Hands-on experience training machine learning models.
• Experience with model training and optimisation, especially improving iteration speed.
• Strong mathematical reasoning.
• Relevant backgrounds may include LLMs, VLMs, diffusion, or large-scale ranking and recommendation; an existing ML job title is not required.
• Must work fully onsite in Bermondsey, London.

Nice to have

• Experience with VLAs or world models.
• Robotics simulation and evaluation, including NVIDIA Isaac Sim.
• Experience with Unitree G1 robots, VLM orchestration, or deploying learned policies on physical hardware.
• Background in mathematics, physics, quantitative fields, or competitive programming.

Culture & Benefits

• First dedicated ML hire with ownership from day one.
• Work directly with robots and use fast experimentation loops.
• Equity included alongside the base salary.

Hiring process

• Apply through the provided application route or send a resume to khorsley@acceler8talent.com.