Creative Technologist, Permanent Contract: Startup AI Content Creation/ Production Platform
Location: Remote
Our client is an AI content solution built for prestige fashion and beauty brands. They blend the speed of AI, the craft of human hands and the control of brand leaders to create leading edge digital content at scale.
They seek a Creative Technologist as a senior hands-on craftsperson who turns each house's creative direction into the prompt systems and custom ComfyUI workflows that produce on-brand imagery at SKU scale, as well as editing ad modifying ComfyUI nodes in Python as needed. You sit at the intersection of art direction and applied AI engineering. You think like a creative; you build like an engineer.
This is a permanent contract role. Compensation will be commensurate with experience. Looking to hire immediately.
Responsibilities:
-Develop modular, parameterized prompt templates that produce consistent, on-brand output across categories, subcategories, and SKU variants
-Build component libraries — reusable lighting setups, set descriptors, styling vocabularies, framing conventions — composable by category requirements
-Translate creative direction into prompt vocabulary. Convert language like "a Milanese afternoon in October" into the technical specificity a model can act on
-Maintain version control on prompt assets the same way engineers maintain code: branches, reviews, regression tests against reference images
-Build production ComfyUI workflows that execute brand-specific prompts at SKU scale ControlNet conditioning, IP-Adapter references, regional prompting, latent compositing, multi-pass refinement
-Integrate brand-trained LoRAs (built by the AI Research Engineer) into workflows correctly:weights, layer targeting, interaction with other conditioning
-Write or modify custom ComfyUI nodes in Python where required
-Build batch processing pipelines that take SKU inputs and produce reviewable output without manual intervention per item
-Run structured experiments to identify which prompt variations, model checkpoints, samplers, and parameter combinations produce the highest-quality output for each category
-Diagnose failure modes — inconsistent hardware, wrong proportions, off-brand color, anatomical errors — and engineer them out at the prompt or workflow layer
-Review output before it moves to retouching. Catch issues early
-Partner with the AI Research Engineer on what to fine-tune versus what to prompt-engineer
-Work with QC and retouchers to understand what AI cannot yet deliver and adjust strategy accordingly
-Document everything. Your work must be legible to the next person who joins the team.
Requirements:
-4+ years working professionally with generative image models, with at least 2 years focused on production output rather than experimentation
-Deep, hands-on fluency with Midjourney, Nano, Stable Diffusion (SDXL and newer), and ComfyUI. You know the parameters, samplers, schedulers, and which combinations work for which problems.
-Strong working knowledge of ControlNet, IP-Adapter, LoRAs, and conditioning techniques.
-Demonstrable experience building production ComfyUI workflows — not just consuming others' workflows.
-Python literacy sufficient to read, modify, and write custom ComfyUI nodes.
-Experience working under brand or client creative direction. Agency, in-house brand, or freelance for premium clients all qualify
Apply Now
Send your resume to jodi@trurecruitment.com.