ML Engineer (f/m/d) - Robotics Foundation Models

ML Engineer (f/m/d) - Robotics Foundation Models

100% Zürich

Hexagon Robotics is a division of Hexagon – a global leader in precision measurement. The division develops humanoid robots for industrial sectors to address labor shortages and accelerate the transition from automation to autonomy. Our first humanoid, AEON, was launched in June 2025 and is already in pilots with five customers.

We are hiring a Machine Learning Engineer to join our Foundation Model team, focused on building, evaluating, and deploying state‑of‑the‑art foundation models for humanoid robotic manipulation. You will work on making modern models (VLA, world action models, diffusion policies) work reliably and efficiently on a real industrial humanoid robot.

Your Mission

• Evaluate and benchmark foundation models (e.g. VLA, world action models, diffusion-based policies) on real robotic manipulation tasks
• Design and run experiments to understand model behavior, data requirements, failure modes, and generalization limits on humanoid hardware
• Develop fine-tuning and adaptation methodologies to specialize large pre-trained models for industrial-grade performance on targeted tasks
• Build and maintain clean evaluation pipelines, benchmarks, and metrics to support rigorous model comparison and selection
• Work closely with MLOps/DataOps and Learning teams to integrate selected models into our end-to-end learning and deployment pipeline
• Collaborate with Robotics and Product teams to ensure model choices align with real-world constraints and customer workflows

Your Skillset

• PhD or MSc in Machine Learning, Robotics, Computer Science, or a related field
• Solid understanding of modern multimodal model architectures and training methods
• Strong hands-on experience with modern deep learning frameworks (e.g. pytorch/jax)
• Experience designing experiments and making data-driven technical decisions

Strong advantage

• Experience with VLAs, world action models, diffusion policies, etc. for robotic manipulation
• Familiarity with efficient fine-tuning techniques (e.g. LoRA, adapters, distillation)
• Experience working with real-world, imperfect datasets and hardware-in-the-loop evaluation
• Publications, open-source contributions, or strong applied research projects in ML and/or robotics

What You’ll Get

• Flexible working hours and a hybrid model for real work-life balance
• Generous vacation: 25–30 days depending on age
• CHF 500 mobility credit for sustainable commuting
• Bonus system & strong pension contributions
• Tailored training & development opportunities
• Relocation support for a smooth start
• Discounts on health, mobility & entertainment
• Team events and a flat hierarchy where your voice counts
• A warm, international culture built on respect and collaboration

Great robots need great people.