Computer Vision & Data Pipeline Engineer

About Us
We are building the world’s most valuable robotics training datasets. Our mission is to capture and process large-scale video data of human actions and environments, then deliver structured datasets to leading robotics companies.

The Role
We’re looking for a Computer Vision / Data Pipeline Engineer to take ownership of our video data pipelines — from raw uploads to structured, client-ready datasets. You’ll work with state-of-the-art tools (classic CV models + modern vision-language models) and annotation platforms to automate, scale, and ensure quality.

Requirements

What You’ll Do

• Build scalable pipelines to process and clean massive amounts of video data.
• Automate annotation workflows using classic CV models + vision-language models (ChatGPT, Gemini, etc.) + human QA.
• Package datasets in industry-standard formats (COCO, AVA, pose/action recognition datasets).
• Maintain high data quality: filtering, deduplication, validation, metadata consistency.
• Collaborate with backend engineers and product teams to deliver client-ready datasets.

What We’re Looking For

• Strong Python skills (OpenCV, PyTorch/TensorFlow, NumPy/Pandas).
• Experience with large-scale video dataset processing.
• Familiarity with annotation tools (Label Studio, CVAT, Supervisely, etc.).
• Knowledge of dataset standards used in CV/robotics (COCO, Pascal VOC, AVA).
• Hands-on experience automating labeling and data QA.
• Familiarity with modern Vision-Language Models (ChatGPT, Claude, Gemini, etc.) and how they can assist in labeling/annotation.

Nice-to-Have

• Background in robotics, autonomous driving, AR/VR, or large-scale data companies.
• Prior experience in data annotation/dataset companies (Scale AI, Appen, Labelbox, Hive).
• Prior startup experience (fast iteration, lean teams).