Build an internal software tool that runs uploaded images through an automated AI pipeline.

We're looking for someone to build an internal AI tool for visualizing brand spaces and interior architecture. We want a system our team runs in-house, not a per-project rendering service. How it should work: we upload photos of a real existing space, brand guidelines, dimensions and a short brief. The system generates on-brand interior concepts based on our actual photos, preserving the real architecture - same walls, windows, doors, proportions and camera angle unless we specifically ask for a change. Multiple views of the same design must stay consistent with each other. What we need: - Controlled image-to-image pipelines with structural conditioning (ComfyUI, ControlNet, IP-Adapter or similar) — not prompt-based generation - Multi-view consistency across a single design - A simple interface a non-technical team member can operate - Handover: working files, documentation and a recorded walkthrough so we're not dependent on you afterwards Architecture or interior visualization experience is a plus, but strong pipeline engineering matters more. Please include: - Your portfolio - One example where you preserved thereal architecture of a space while changing the design - A brief note on how you did it The listed budget is a placeholder - we'd like your recommendation on scope and cost.