AI / Computer Vision Engineer — Open-Source Image AI Stack

We’re looking for an experienced AI / Computer Vision Engineer to help us replace several third-party AI APIs with an in-house, self-hosted stack using open-source models. We operate a marketplace for secondhand and vintage clothing. Our current AI pipeline is already producing excellent results, but we want to bring it in-house to significantly reduce our cost per item. Current Pipeline Today, every product goes through: * **Background Removal** — BackgroundErase.com to cleanly remove backgrounds from clothing photos. * **Product Analysis** — Gemini Flash Lite analyzes product images and generates the title, description, product attributes, and condition/defect analysis. * **AI Flat-Lay Generation** — Gemini Flash removes the hanger and transforms the original photo into a clean, symmetrical flat-lay while accurately preserving the actual garment, graphics, colors, tags, stains, and defects. What We Need We want someone who can identify the best **open-source models**, build a proof of concept, and ultimately deploy them on our own GPU infrastructure. Quality is critical. Our existing pipeline is dialed in, so the new solution must match or exceed our current results while reducing inference costs. Ideal experience includes **computer vision, VLMs, image segmentation, diffusion/image-to-image models, Hugging Face/PyTorch, GPU inference optimization, and deploying self-hosted AI models at scale.** Experience with **fashion, apparel, ecommerce product photography, or garment image processing** is a major plus. When applying, please share relevant projects you've built and **which open-source models/approaches you would initially consider for replacing our current stack.**