Applied Physical AI Lead

Applied Physical AI Lead - Shadow Robot

Gospel Oak, London | 3-6 month contract, PAYE or Ltd | Start ASAP | [Day rate - to confirm]

Have you put learned policies on real robots - and want to be the person who decides what an experienced robotics team builds next?

Would you rather spend the next few months driving real decisions than writing another research proposal?

Want to see what you can do with 30 years of dexterous hardware behind you?

Shadow Robot builds some of the most advanced dexterous robotic hands in the world. Our hands are used by research labs and frontier AI companies, including Google DeepMind and OpenAI. We are an employee-owned team of around 35 people, from 17 nationalities, based a 2-minute walk from Gospel Oak station.

Our engineers know the hardware inside out. What we want now is an applied AI lead who can set the direction on the learning side, make the calls, and get working systems onto our robots quickly.

What you will do

• Set the technical direction for robot learning and applied AI at Shadow - which models, which approaches, and why
• Turn that direction into working systems on real dexterous hardware, using the best tools available rather than starting from a blank page
• Lead and unblock a team of experienced engineers, so decisions are made and work moves forward
• Shape the data, training and evaluation approach that our proofs of concept run on
• Leave behind a clear plan, and the foundations of a permanent role, at the end of the contract

What you will bring

• Hands-on experience deploying learned behaviours on physical robots - for example imitation learning, reinforcement learning, VLA or foundation models, sim-to-real
• Evidence of leading people or a technical area, not only contributing to one
• An applied mindset: you pick the right existing tool, build it, test it, and change course when the data says so
• The judgement and confidence to make a call and explain it clearly to senior engineers
• Strong Python, and comfort working in a Linux and ROS-based robotics stack

Extra points for

• Manipulation or contact-rich tasks, especially with hands or grippers
• Experience in a startup or small team where you had to deliver with limited resources
• Data collection and teleoperation pipelines for robot learning

Why this is worth your time

• Real ownership from week one, on hardware few people ever get to work with
• A short, defined engagement with a genuine route to a permanent leadership role if we are the right fit for each other
• A small, friendly, low-ego team: we care about each other, we enjoy challenging the impossible, and we are compassionate

This is not for you if

• Your experience is mostly in simulation, academic research or LLM applications, without shipping on physical robots
• You want a long ramp-up before you take ownership

The practicals

• Contract: 3 months with an option to extend, or up to 6 months. PAYE or through your limited company 
• Rate: Negotiable
• Location: On-site in Gospel Oak, London 
• Start: As soon as you are available
• Notes: We cannot offer visa sponsorship for this contract. Not suitable for recent graduates

How to apply: Apply here and Helen, our hiring partner, will be in touch within two working days. 

No agencies, please.