Robotics ML Scientist: Build Large-Scale Foundation Models

Enact; Zürich; On-site; CHF 180,000 - 240,000; Full time

Benefits offered by this job

Visa sponsorship

Relocation benefits

Equity participation

On-site Zurich setup

Job summary

Enact in Zürich is seeking a researcher to build the base intelligence layer for robotics, training large-scale force-aware foundation models to enable dexterous manipulation across tasks and environments.

You will design and run pretraining, own end-to-end experiments, and create data mixes and evaluation sets to push AI-driven manipulation from pixels to actions, with real-world and simulated testing.

Qualifications

• Deep experience developing large-scale predictive or generative models, including transformer or diffusion architectures.
• Strong probability, statistics, and ML fundamentals; ability to design rigorous experiments.
• PhD or equivalent research experience with a track record in ML, CV, robotics and related fields.
• Proficiency in ML computing frameworks like PyTorch or JAX, optimizing for performance at scale.

Responsibilities

• Design and execute large-scale pretraining runs for robot foundation models, defining architectures, objectives, and training curricula.
• Own experiments end-to-end, from data specification through training to real-robot evaluation, to understand scaling laws and architecture tradeoffs.
• Build datamixes and evaluation sets to train and assess robots on novel tasks across real-world and simulated settings.
• Research methods for improving sample efficiency and robustness for reliable real-world deployment.

Skills

Large-scale models

Probability & ML fundamentals

Experiment design

PyTorch/JAX

Robotics deployment motivation

Education

PhD or equivalent

Tools

PyTorch

JAX

Job description

Enact in Zürich is seeking a researcher to build the base intelligence layer for robotics, training large-scale force-aware foundation models to enable dexterous manipulation across tasks and environments.

You will design and run pretraining, own end-to-end experiments, and create data mixes and evaluation sets to push AI-driven manipulation from pixels to actions, with real-world and simulated testing.