Computer Vision Engineer — Warehouse Robotics

Kora Intelligence

Hybrid (Pittsburgh, US) · $112K – $136K · Full-time · Posted 6 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

Perception is the thing holding the next robot release back. Not in a dramatic way, just in the real engineering way: reflective packaging, awkward shelf angles, occlusion, and edge compute limits all adding up to a system that is close but not ready. We need a vision engineer who likes that kind of problem. You will spend time profiling failures, working with robotics and hardware teams, and improving localisation performance where it actually matters on the warehouse floor.

Responsibilities

Improve object detection and localisation performance for our next-generation warehouse robots
Profile failure cases across lighting, reflective surfaces, occlusion, and shelf geometry
Work with robotics and hardware teams to improve the end-to-end perception pipeline, not just the model checkpoint
Design evaluation workflows tied to operational metrics such as missed picks and recovery time
Document model and data decisions clearly for the broader robotics engineering group

Requirements

4+ years of computer vision engineering in robotics, industrial automation, or production vision systems
PyTorch for training and evaluating detection or segmentation models
Computer Vision fundamentals including calibration, localisation, detection, and error analysis under real-world constraints
NumPy for data handling, transformation, and profiling workflows
Docker for reproducible inference and training environments
Experience balancing model quality against edge-device latency and memory constraints