The Role
Co-lead AI research focused on multimodal video-action foundation models for dexterous robotics. Responsibilities include model architecture and large-scale training, end-to-end data pipeline development, real-world robot evaluation, deployment, and mentoring AI researchers and engineers. The role requires deep generative modeling expertise, distributed PyTorch development, multi-GPU pretraining experience, and collaboration with hardware and robotics teams.
Summary Generated by Built In
mimic robotics is a frontier physical AI company building the full stack of general-purpose dexterous manipulation. Founded in 2024 by ETH Zurich researchers, mimic pioneered Video-Action Models for robotics while also developing in-house dexterous hand hardware. Our growing team brings together world-class researchers working across our offices in Zurich and San Francisco.
We are hiring an AI Research Engineering Lead in Zurich. You will co-lead the AI team, focusing your efforts on pushing the frontier of multi-modal, video-action models (VAMs) targeting higher generalisation capabilities and improving deployment to real world industrial tasks.
Key Responsibilities
- Drive Training & Architecture: Lead research, architecture design, and training for large-scale, multi-task foundation models across pre- and post-training.
- Scale the Data Flywheel: Manage the end-to-end data pipeline—from raw human teleoperation data specification to model training, fine-tuning, and real-world robot evals.
- Lead & Mentor: Guide AI research engineers and scientists, setting high standards for ML engineering, experimental velocity, and empirical evaluation.
- Cross-Functional Deployment: Partner with hardware and robotics teams to align data collection efforts with benchmarking of trained models and stable real-world deployment.
Requirements & Qualifications
What You Bring:Research Background: PhD in CS/ML or equivalent PhD-level industry research experience (e.g., Senior/Staff AI Scientist at a frontier lab). We heavily prioritize generalist ML and deep pre-training over pure robotics or classical control/RL background.
Generative Modeling at Scale: 3+ years directly building and scaling generative architectures (diffusion, flow matching, autoregressive models, VLMs, or video models).
Distributed Engineering: 4+ years of hands-on PyTorch (or similar) development, including multi-GPU pre-training setups (FSDP, DeepSpeed, Megatron) and scaling law mechanics.
Technical Leadership: Proven track record of mentoring junior engineers, supervising PhD interns, or leading collaborative research projects.
Location: Based in or willing to relocate to Zurich for full-time, in-person work (San Francisco based candidates will also be considered).
Out of Scope: This is a deep model-building role. Candidates limited to classical control/RL without generative models, or application-level RAG/API wrappers, will not be considered.
- Competitive Package: Base salary + substantial equity stake.
- Frontier Impact: Direct ownership of core AI strategy alongside world-class talent in a flat, fast-moving environment.
- Resources & Perks: Modern GPU compute, hardware test benches, free gym/sports subscriptions, team meals, and full relocation support.
You will have the opportunity to shape a robotics & AI company from the ground up. In flat hierarchies you will work directly with the founders and some of the best talents in the robotics space. At mimic, you are accepted for who you are. As part of our ongoing journey towards creating a diverse and inclusive environment we encourage everyone to apply and we are looking forward to you bringing along your knowledge, personal experiences, and fresh perspectives. Together we can solve some of the greatest challenges in robotics.
Skills Required
- PhD in Computer Science, Machine Learning, or equivalent PhD-level industry research experience
- 3+ years building and scaling generative architectures such as diffusion, flow matching, autoregressive, vision-language, or video models
- 4+ years of hands-on PyTorch or similar development
- Experience with multi-GPU pretraining setups, including FSDP, DeepSpeed, or Megatron
- Experience with scaling law mechanics
- Track record of mentoring junior engineers, supervising PhD interns, or leading collaborative research projects
- Based in or willing to relocate to Zurich for full-time, in-person work; San Francisco candidates may also be considered
- Deep model-building experience; classical control or reinforcement learning alone is insufficient
Am I A Good Fit?
Get Personalized Job Insights.
Our AI-powered fit analysis compares your resume with a job listing so you know if your skills & experience align.
Success! Refresh the page to see how your skills align with this role.
The Company
What We Do
Physical AI to scale your most tedious tasks from manufacturing to logistics. Our robots intuitively learn new skills from you and operate autonomously in any environment.
