Jared G.

ComfyUI Developer - Product Imagery, Custom Models & AI Video

I build generative AI image and video pipelines in ComfyUI — the kind of work where "it has to look the same in every shot" actually gets solved instead of hand-waved. What I do: Product and brand imagery — your product placed into scenes that look photographed, with the label and packaging kept pixel-for-pixel identical to the file you send me. No redrawing your product, no invented text on the box. Original character pipelines — custom models for Flux and WAN 2.2 that hold one invented character steady across a full set of images and video, from dataset curation through training to finished output. These are original characters I create, not real people. ComfyUI workflows — text-to-image, image-to-video (WAN 2.2), restoration and upscaling (CodeFormer, RealESRGAN), and frame interpolation (RIFE). I build workflows that run reliably, and I document them so you can run them without me. Cloud GPU deployment — I set up and run training and generation workloads on RunPod (A100, RTX 6000 Ada), so you don't need your own hardware. Pixel art and game assets — AI-assisted sprite and animation sheet generation for game projects. One thing I don't take on: work built around a real person's likeness. My character work is original characters only. I came to AI from 18 years in the structural trades, where "close enough" gets people hurt. I bring that same standard here: I test before I deliver, I tell you honestly what the tech can and can't do, and I don't disappear mid-project. If your AI pipeline is broken, or you need one built, send me the details and I'll tell you straight whether I can fix it.