Role: AI Developer – LoRA (Characters & Garments)
Location: Remote (location-independent)
About the Role
We’re kicking off the DHTA AI Studio and are looking for an experienced AI Developer to help us build meaningful technology at the intersection of AI, ethics, and sport.
This role focuses on designing, training, and optimizing high-quality LoRA models for human characters (athletes) and garments (jerseys, shorts, kits, accessories). You’ll work on building modular, reusable LoRAs that can be reliably combined (multiple characters + multiple garments) within the same generation pipeline.
Responsibilities
* Own the end-to-end LoRA pipeline for identity and apparel: dataset preparation, training, evaluation, inference, and workflow integration
* Train and fine-tune LoRAs with strong visual consistency and controllability (rank/alpha, schedulers, regularization and negative sets, captioning strategies), managing overfitting, bleed, and editability
* Develop workflows that support multi-LoRA composition (character + garments), including adapter weighting, masking, prompt/caption alignment, and conflict mitigation
* Build and customize AI pipelines using ComfyUI (or equivalent node-based systems)
* Write and maintain custom nodes, scripts, and automation in Python / PyTorch
* Optimize datasets, training parameters, and inference pipelines for repeatable, high-quality outputs
* Optimize inference performance (VRAM usage, batching, caching) and define quality metrics and validation sets
Required Skills
* Strong experience with Python and PyTorch
* Deep understanding of LoRA training for image and/or video generation
* Hands-on expertise with ComfyUI and custom workflow design
* Solid knowledge of diffusion models and conditioning techniques
* Ability to work independently in a remote, fast-paced startup environment
* Strong problem-solving skills and attention to detail
Tech Stack
Python, PyTorch, diffusers / Stable Diffusion tooling, LoRA training frameworks, ComfyUI internals, dataset tooling, GPU optimization.
To Apply
Send your CV or a short intro outlining your relevant experience to
* 📩 hiring@dhta.ai