AI Video/ComfyUI Developer

Company: BoredSexy / UTON

Location: Kuala Lumpur preferred, remote considered

Type: Contract or full-time

Compensation: Based on experience

We are building an internal AI video production stack for character-consistent synthetic media.

We need a hands-on AI developer who can help us move beyond prompt-only generation and build a repeatable pipeline for training, testing, and deploying custom character models, LoRAs, ComfyUI workflows, and open-source video generation models.

This role is technical, practical, and production-focused. We are not looking for someone who only follows tutorials. We need someone who can experiment, debug, document, and turn messy AI workflows into internal tools our content team can use every day.

• Build and maintain ComfyUI workflows for image and video generation
• Train, test, and manage custom LoRAs for consistent AI characters
• Set up repeatable workflows for reference images, pose control, identity consistency, and prompt templates
• Evaluate video generation models, including Wan / Alibaba Wan, HunyuanVideo, Seedance-style pipelines, and related open-source tools
• Build internal tools for batch generation, queueing, checkpoint management, prompt libraries, and asset review
• Help create a clean production pipeline from concept to reference image to generation to QA to storage
• Run experiments on cloud GPU infrastructure such as RunPod, Vast.ai, Lambda, or equivalent
• Document workflows so non-technical team members can use them safely
• Work with the creative team to improve visual consistency, realism, and output quality

• Strong Python experience
• Real experience with ComfyUI, Stable Diffusion, Flux, SDXL, or similar generation pipelines
• Experience training or fine-tuning LoRAs, DreamBooth, textual inversions, or related adapters
• Comfort working with custom nodes, model weights, checkpoints, VAEs, ControlNet, IPAdapter, FaceID, pose/reference workflows, and video-generation extensions
• Basic cloud GPU experience
• Ability to debug broken environments, dependency conflicts, CUDA issues, and model-loading errors
• Good judgment around data handling, model licensing, privacy, and content compliance
• Clear communication and documentation habits

• Experience with Wan / Alibaba Wan video models
• Experience with NSFW-safe infrastructure, content moderation, or age-gated creator platforms
• Experience building internal creative tools, dashboards, or simple UIs
• Familiarity with Google Drive, Sheets, APIs, and automation workflows
• Experience with video post-processing, upscaling, frame interpolation, captioning, or batch rendering
• Experience with LoRA dataset preparation, tagging, captioning, and quality control

Some projects involve mature, adult-oriented, age-gated creative media. We only work with consenting adults, synthetic characters, and lawful content. The role requires professional maturity, privacy awareness, and good judgment.

If you are uncomfortable working near mature-content workflows, this role is not a fit.

In the first 30 to 60 days, you should be able to:

• Set up a stable ComfyUI environment for internal use
• Train or fine-tune at least one usable character LoRA
• Build a repeatable generation workflow for a specific character
• Document the workflow clearly enough for a non-engineer to run it
• Recommend a practical path for internal AI video generation using Wan or similar models
• Identify the biggest blockers around quality, compute cost, consistency, and compliance

Send:

• A short note about your AI image/video experience
• Links to relevant GitHub repos, demos, workflows, or screenshots
• A description of the most complex ComfyUI or LoRA setup you have built
• Your expected rate or salary range
• Your availability and location