Embodied AI Foundation Models Engineer

Genesis AI; Paris; On-site; EUR 90,000 - 150,000; Full time

Benefits offered by this job

Vision-Language-Action models exposure

Job summary

Genesis AI in Paris is seeking a senior ML researcher to build end-to-end models for robotics control, covering data curation, evaluation, architecture, training and inference stacks.

You will collaborate with simulation and robotics teams to curate diverse, large-scale datasets and push the boundaries of embodied perception and generative simulation.

A strong publication record or impactful projects and 5+ years of experience, plus deep Python expertise, are highly valued.

Qualifications

• Candidate must have strong track record in foundation model research and engineering.
• Shows ownership and ability to solve problems independently.
• 5+ years of relevant ML research or engineering experience with publications or impactful projects.
• Focus on high-quality data and reproducible experimentation.
• Production-level Python proficiency.

Responsibilities

• Build end-to-end ML models for robotics control, incl. data curation, evaluation, and training/inference stacks.
• Collaborate with simulation and robotics teams to curate diverse, large-scale datasets.
• Develop generative simulation methods for scalable data and models to generate tasks and rewards.
• Work with a team focused on advancing general-purpose Physical AI.

Skills

Full-stack foundation model research

Ownership and initiative

ML research

Data and code quality

Python programming

Job description

Genesis AI in Paris is seeking a senior ML researcher to build end-to-end models for robotics control, covering data curation, evaluation, architecture, training and inference stacks.

You will collaborate with simulation and robotics teams to curate diverse, large-scale datasets and push the boundaries of embodied perception and generative simulation.

A strong publication record or impactful projects and 5+ years of experience, plus deep Python expertise, are highly valued.