AI Virtual Try-On (VTON) / Stable Diffusion Engineer
AI Virtual Try-On (VTON) / Stable Diffusion Engineer
Freelance / Contract | Remote | Fashion & E-Commerce AI
About Us
We’re a content production company working at the intersection of fashion, technology, and creative media. Our clients expect photorealistic, high-quality imagery suitable for luxury e-commerce catalogs. We’re now building an AI-driven virtual try-on workflow that converts ghost mannequin product shots into model images with consistent faces, poses, and catalog-ready quality.
Role Overview
We’re looking for a Stable Diffusion / ComfyUI engineer with experience in ControlNet, IP-Adapter/InstantID, and virtual try-on (VTON) pipelines to help us set up a repeatable system.
Responsibilities
Design and implement a ComfyUI (or AUTOMATIC1111) pipeline for ghost mannequin → on-model try-on.
Integrate ControlNet (pose/depth) for consistent body/angles (front/side/back).
Apply IP-Adapter or InstantID for model face/hair consistency across all garments.
Configure garment conditioning / VTON modules (e.g., IDM-VTON, GP-VTON) for accurate drape and print fidelity.
Ensure outputs meet luxury e-commerce quality: neutral backgrounds, sharp garment edges, accurate fabric textures.
Document the pipeline for internal use and train our team on running batches.
Optional: Assist with post-processing automation (upscaling, seam cleanup, relighting).
Requirements
Proven experience with Stable Diffusion, ComfyUI, or similar frameworks.
Strong knowledge of ControlNet (pose/depth) and IP-Adapter / InstantID.
Hands-on with virtual try-on / fashion-focused diffusion models.
Portfolio or examples of photorealistic AI model/garment outputs.
Ability to deliver repeatable, production-ready pipelines (not just one-off generations).
Strong communication and documentation skills.
Nice to Have
Experience with LoRA / custom training for garment fit or fabric fidelity.
Familiarity with Python automation for batch jobs.
Past work with fashion e-commerce imagery.
What We Offer
Freelance/contract engagement with potential for ongoing collaboration.
Competitive project-based compensation ($3K–10K depending on scope and experience).
Opportunity to shape an AI-first fashion production pipeline at scale.
Flexible, remote-first collaboration with a fast-moving creative team.
How to Apply
Please submit:
Your portfolio or sample outputs (ideally fashion/virtual try-on related).
A short note describing your experience with ControlNet + IP-Adapter and any VTON models you’ve worked with.
Your availability and estimated project rate.