Team Lead Machine Learning - Physical AI

About the role

A Zurich-based company is hiring a Team Lead Physical AI (f/m/d) to lead a team that turns human demonstrations into reliable manipulation skills on its humanoid platform. The role owns the path from teleoperated demonstration to deployed policy: how data is collected, what it is trained on, and how it is proven on real robots in real customer environments. It is a leadership role with a hands-on core: you will hire, mentor and manage the team and own its roadmap while staying technically involved across the imitation learning stack.

Key responsibilities

• Build and lead a team of ML engineers and researchers, set a culture of rapid experimentation with rigorous evaluation, and keep the roadmap honest about research versus delivery.
• Lead the technical development of visuomotor policies: build the roadmap, prioritize directions, review experiments and code, unblock the team and take on hard problems yourself.
• Set the bar for evaluation by building real-robot and simulation protocols with enough statistical rigor to support go/no-go decisions.
• Establish regression testing so improvements are measurable and failures are diagnosable.
• Ship policies into production with Product, MLOps, Controls and Hardware teams, and close the loop on field failures.
• Co-engineer with strategic partners and help build new partnerships and academic collaborations.

Required profile

• Deep, hands-on expertise in ML research and development (PhD, MSc or equivalent industry experience).
• 2+ years managing or technically leading a team, with the appetite to keep doing both people and technical work.
• Track record of translating research into working systems (papers and shipped product or demo).
• Comfort in a fast-paced startup-like environment inside a larger organization.
• Strong advantage: publications or visible contributions in imitation learning, embodied AI or robot learning.

Required skills

• Professional experience training and deploying ML models, with strong Python and deep learning framework skills (PyTorch).
• Practical experience managing and scaling datasets.
• Advantage: imitation learning and reinforcement learning for robotic manipulation, humanoid, bimanual or mobile platforms, teleoperation systems, diffusion and flow-matching policies, action chunking, VLA architectures, distributed training and experiment tracking.

What we offer

• Flexible working hours and a hybrid model.
• 25-30 days of vacation depending on age.
• CHF 500 mobility credit and a bonus system with strong pension contributions.
• Tailored training and development, relocation support, discounts on health, mobility and entertainment, and team events.