Expert AI Engineer — Image Understanding & Model Fine-Tuning

We sell personalised artwork. A customer uploads a photo of themselves or their own pet as a reference, and we generate a stylized painted portrait — which we print and ship as a canvas. We need an expert AI engineer for two things: 1. Image understanding. Before generating, we need to reliably read the reference photo — person or pet, number of subjects, breed, framing, image quality — so we route to the right template and prompt. This step is currently unreliable and we generate from the wrong prompt too often. We want a proper evaluation set and measured accuracy, not tuning by feel. 2. Our own models. We currently depend on third-party generation APIs. We want to move to models we fine-tune and host ourselves. You'd pick the base models, curate training data from our existing work, run the fine-tuning, evaluate results properly, and deploy. Requirements - Hands-on experience fine-tuning open-weight image generation models — LoRA, full fine-tunes, dataset curation - PyTorch, GPU infrastructure, inference optimisation - Solid software engineering. Our stack is TypeScript/Node - Shipped generative image systems to production - Clear written English, async-friendly To apply: Skip the generic cover letter. Answer these: - Which base model would you fine-tune on our art styles, and how would you approach it? - How would you build the image-understanding step, and how would you prove it's accurate? - Link one production system you've built involving generative image models.