Job Description

 

Our client is a seed stage physical AI company based in Paris, spun out of one of Europe's leading robotics research institutes. Their software lets an operator show a humanoid robot a task a few times and have it working on the production line in under an hour, with no code.

 

The software is hardware agnostic and runs across multiple humanoid platforms. It is already deployed with major European manufacturers in automotive, industrial equipment and consumer goods.

 

The team is small and research led, with a world renowned robotics and machine learning researcher as CTO. It combines the rigour of a research lab with the pace of an early stage startup.

 

The role

 

You'll build the perception layer that lets the robots understand the world well enough to act in it. Factories are tough visual environments: cluttered scenes, shifting light, objects the robot has never seen before, and nobody on hand to retune things. Your methods need to work out of the box, on every site, with minimal data.

 

This is research that ships. What you build goes straight into the product and runs on real robots in real factories. As one of the first 15 hires, you'll also help set the company's research direction.

 

Responsibilities

 

- Design, prototype and evaluate perception for robotic manipulation: object recognition, segmentation, pose estimation and scene understanding

- Push the state of the art in robust, data efficient perception for embodied AI

- Run experiments on real robots and help build the data and evaluation tools behind them

- Turn research prototypes into perception components ready for production

- Work closely with the Robot Learning and Integration teams to get research deployed

- Help shape the long term research roadmap

 

Requirements

 

- PhD or MSc in computer vision, robotics, machine learning or a related field

- Background at a leading research lab, academic or industry

- Publications at top venues such as CVPR, ICCV, ECCV, NeurIPS, ICML, CoRL, RSS or ICRA

- Deep expertise in computer vision or machine perception: recognition, segmentation, pose estimation, scene understanding or 3D perception

- Strong Python and PyTorch

- Hands on experience training and debugging modern vision systems

- Professional English

 

Nice to have

 

- 3D vision: depth estimation, point clouds, geometric reasoning or scene reconstruction

- Perception for manipulation, including grasping

- Experience with real robots (humanoids, arms or mobile bases). If your background is pure vision, they'll help you ramp up on robotics

- Postdoc or industry research lab experience

- A track record of taking research into deployed systems

 

What's on offer

 

- Open salary for the right candidate

- Early and generous founding equity

- Full relocation support to Paris

- EU and non EU candidates welcome

- English speaking team, no French required

- A fleet of advanced robots, including humanoids, to run your research on

- Real ownership of the perception roadmap

 

 

Apply here or message Cal at Edison Smart directly for a confidential conversation. All applications are handled in strict confidence.