Founding Machine Learning Engineer

Founding Machine Learning Engineer

Paris, France · Hybrid · Full-time

About the company

Operate Talent is recruiting on behalf of one of the fastest-growing AI data companies, backed by top-tier VCs. They provide real-world training data to the world’s leading AI and robotics labs, with a mission to advance embodied AI through real-world data and research.

About the role

As Founding ML Engineer, you will own the video data pipeline end to end, from raw capture in the field to validated datasets that frontier robotics labs train on. You will work directly with the founders and with some of the biggest AI labs, on live production with real volume and hard deadlines.

What you’ll own

• Petabyte-scale multi-camera video (RGB-D, RGB, IMU, metric depth, hardware sync), turned from raw bundles into validated episodes
• Egocentric video with multilingual narration: ASR, translation, chapter segmentation, and QA against customer taxonomies
• Quality control as an ML problem: benchmarking, fine-tuning and distilling VLM judges, and owning evals and thresholds
• Scalable QC/QA pipelines that ensure no clip ever ships out of spec
• Annotation at scale, including hand tracking and fine-grained manipulation
• Next modalities such as tactile and teleoperation

Requirements

• 3+ years of applied ML in computer vision or robotics, ideally with a top engineering school background. Outliers with less experience are welcome: the bar is what you’ve built
• Hands-on experience running VLMs as judges in production, including calibration, fine-tuning and distillation
• CV/robotics models deployed in production: detection, tracking, pose/hand estimation, depth, VIO, action recognition
• Deep knowledge of the video stack: codecs, frame timing, multi-view geometry, calibration, sensor alignment
• Exceptional achievements such as competition rankings, hackathon wins, or projects at real scale
• An active GitHub and Hugging Face presence
• Agentic engineering as a craft
• EU residency (visa sponsorship is not available)

Nice to have

Egocentric vision, IMUs, MCAP, ROS, world models, VLAs or robot foundation models

Tech stack

PyTorch, CUDA, multi-GPU training, vLLM on H100s, HEVC + IMU via MCAP, Postgres, S3, Kubernetes

What’s on offer

• A very competitive salary plus meaningful equity
• Your own GPU budget
• A founding role working directly with the founders and the biggest AI labs
• A new Paris office opening soon, with a hybrid setup

Hiring process

1. Short application
2. Technical conversation
3. Build task on real pipeline data
4. Final conversation with a founder

You’ll get an answer within 48 hours.

How to apply

Email hr@operatetalent.com with the subject line “Founding ML Engineer – [Your Name]” and

include:

1. Your CV
2. Your GitHub link
3. Your Hugging Face link