We build smart camera systems for retail stores — the kind that track inventory levels, detect shelf gaps, and flag planogram compliance in real time. We're 22 people, growing fast, and deeply excited about what computer vision can do when it's applied to messy, imperfect real-world environments. We need a junior CV engineer who has worked with image datasets before, knows PyTorch and at least one computer vision framework, and is genuinely energised by the challenge of getting models to work in production conditions — bad lighting, partially obscured products, cameras at weird angles. Our computer vision lead has eight years in the field and will actively mentor you. This is not a solo job. You'll have daily guidance and a clear growth path.
Responsibilities
Train and evaluate computer vision models for shelf monitoring tasks
Prepare and augment training datasets from store camera footage
Test model performance across diverse lighting and angle conditions
Document experiments and maintain a model performance log
Collaborate with the hardware team on camera calibration and deployment constraints
Requirements
Working knowledge of PyTorch for image model training and evaluation