I just open sourced a year of work and I don't have the experience running open source projects...

Over the past year I have been building an end to end pipeline for fine tuning, and in doing so, I accidentally came across a problem that apparently not many people are interested in solving, so diverged to build a solution I genuinely believe solves a real problem — and now I'm at a fork in the road.

The problem we're solving:

Industries like automotive, healthcare, legal, and finance are sitting out of the AI wave — not because they lack data, but because their data is too sensitive to use directly for fine-tuning. Nobody wants to feed real customer conversations into a training pipeline and hope for the best.

What UNCASE does:

UNCASE (Unbiased Neutral Convention for Agnostic Seed Engineering) is a framework for generating high-quality, privacy-safe synthetic training data for LoRA/QLoRA fine-tuning. It's designed to be industry-agnostic, reproducible, and auditable — so regulated industries can actually participate in building their own AI without compromising their data.

We've already used it to fine-tune a production sales assistant for an automotive dealership group with unprecedented results. The pipeline generates, scales, and evaluates synthetic conversations with tool-calling support, in ChatML/Hermes format, using tiered model orchestration.

Why I'm posting:

I have zero experience with GitHub best practices, PR workflows, contribution guidelines, or community building — and I don't think I'm the right person to own that side of this.

I truly and genuinely believe that this could be something big due to the interest it has spiked on the few business people I've talked to about it, but just recently decided to open source it, and for that I need experts and high drive actors that get involved with the project. If you're someone who:

• Finds the problem space genuinely interesting (AI + regulated industries + synthetic data)

• Knows how to run an open-source project the right way

• Wants to be an early, meaningful contributor to something with real traction

...I'd love to talk. This isn't a job posting — it's an invitation to co-own the direction of the project from the community side.

👉 github.com/uncase-ai/uncase -- take a look and tell me what you think. (This made Astrapp.lat actually be useful, which was the original project I started out building a little over a year ago).