Computer Vision Engineer — Retail Visual Search

Feature: photo search. User takes a picture of a jacket they see on the street. App shows them the closest matches from our 2 million product catalogue in under 200ms. This is the #1 requested feature from our users for the past 18 months. We've been putting it off because we didn't have the CV expertise in-house. Now we're hiring for it. The core work: image embedding pipeline, similarity model fine-tuned on our catalogue, integration into our existing product API (Python/FastAPI, hosted on AWS). We already have a pre-annotation pipeline for product images — you'd be building on top of that. We're not looking for someone to recommend a third-party API. We've evaluated those options. We want it built properly.