Computer Vision Engineer

Snapshot

The mission of 120ft Factory is to start the new western industrial revolution. We are building the essential infrastructure to close the "Physical AI Deployment Gap." To make agentic workflows successful in the physical world, our AI must have perfect, real-time ground truth. That requires a flawless observability layer.

Description

As the World View & Data Flywheel Engineer, you work on the critical observability layer of our architecture. You are the bridge between the 'reality of atoms' and our AI models. Your mission is to extract raw telemetry, 2D/3D vision data, and physical states from heterogeneous industrial hardware, translating it into the 'Safe-State Projections' our deterministic supervisor needs, and the 'World View Video Logs' our VLAs learn from. You will vertically integrate the system, building custom robot integration packages that feed a continuous Reinforcement Learning (RL) flywheel, pushing agentic capabilities directly to the shop floor.

Responsibilities

• Own the 'World View' observability layer: architect the pipelines that capture real-time telemetry, 3D point-cloud data, and video logs from active robot cells.

• Develop 'Safe-State Projections' by fusing raw data from industrial vision systems, PLCs, and robot controllers into the structured variables required by our deterministic supervisor.

• Design and deploy custom Robot Integration Packages to extract high-fidelity data directly from heterogeneous hardware.

• Train, deploy, and manage deep learning models (object detection, instance segmentation) that give our systems spatial awareness and semantic scene understanding.

• Architect the end-to-end RL-Flywheel, ensuring every robot success, failure, and edge-case is automatically captured, annotated, and fed back into foundational World Models.

• Work closely with the Systems Engineers and VLA Engineers to vertically integrate agentic workflows, ensuring hardware data seamlessly fuels cloud-side generative models.

Requirements

• Master's or Bachelor's degree in Mechatronics, Robotics, Computer Science, Computer Vision or equivalent practical experience.

• Deep, hands-on experience integrating and developing solutions for industrial robotics (e.g., ABB, Universal Robots, Yaskawa) and PLC synchronization.

• Strong background in Machine Vision (2D/3D) and deploying deep learning models (object detection, instance segmentation) in production environments.

• Exceptional programming skills in Python and C++, with a track record of building custom add-ons, URCaps, or hardware drivers.

Preferred Qualifications

• You understand that an AI is only as good as its ground truth; you are obsessed with robust sensor fusion and flawless data extraction from messy, real-world environments.