Generative AI Training for Creative Production

Looking for Freelance Generative AI Data Scientist / AI Model Training Expert

I am looking for a freelance Data Scientist / AI Model Training Expert who has strong hands-on experience with LoRA training for images and videos.

The work is for an AI creative production platform focused on image generation, video generation, character consistency, image-to-video workflows, and animation-style production.

What I Need

I need someone who can:

* Train image LoRAs for characters, styles, products, costumes, props, and environments.
* Train or fine-tune video LoRAs / motion LoRAs for movement, camera motion, action, and frame consistency.
* Work on image-to-video and video-to-video workflows.
* Prepare, clean, caption, tag, and organize image/video datasets.
* Improve character consistency across different poses, scenes, outfits, and camera angles.
* Reduce issues like identity drift, flickering, bad anatomy, unstable backgrounds, warping, and poor motion.
* Test different training settings and improve output quality.
* Build a repeatable workflow that can be used for multiple characters and client projects.

Required Experience

You must have practical experience with:

* LoRA training
* Image-generation model fine-tuning
* Video-generation workflows
* Character LoRA / style LoRA / product LoRA / motion LoRA
* Dataset preparation for image and video training
* Captioning, trigger words, overfitting, underfitting, and model quality testing
* Python and machine learning
* GPU-based model training

Preferred Tools / Tech Knowledge

Experience with any of the following is preferred:

* ComfyUI
* Kohya training workflows
* Diffusion-based image/video models
* Image-to-video pipelines
* Video-to-video pipelines
* Pose-guided generation
* Face-consistency workflows
* Reference-based generation
* Python
* PyTorch
* Docker
* Git
* Cloud GPU setup

Important

This is not a prompt-writing job.

This is not a traditional dashboard/data analytics job.

I need someone who has actually trained models, tested outputs, fixed poor results, and understands how to make AI-generated characters, styles, products, and scenes stay consistent across both images and videos.

Please Apply With

* Examples of LoRA models you have trained.
* Before/after results.
* Image or video generation work you improved.
* Your training process.
* Tools and models you have worked with.
* Your hourly rate or project-based pricing.
* Your availability.