Most AI projects don't fail at the model. They fail at the data underneath it — pipelines nobody trusts, training sets nobody can reproduce, and agents wired into systems that were never designed to be queried. That's the layer I work on.
Services
Narrow on purpose. These are the engagements where deep warehouse experience makes the difference between a demo and something that survives contact with production.
The unglamorous work that decides whether a fine-tune is worth running: sourcing, cleaning, deduplicating, and versioning the data that goes into a model.
Agents that carry real operational load — intake, scheduling, routing, follow-up — rather than answering questions about a FAQ page.
How the work goes
Engagements usually run in this order, because skipping the first step is what makes the third one fail.
A short, paid discovery on your real systems — schemas, volumes, quality, and the gap between what the data is assumed to contain and what it contains.
Reproducible movement and transformation with tests, monitoring, and a clear failure story. If it can't be re-run from scratch, it isn't finished.
Fine-tuning, evaluation, or an agent wired into the systems that matter — built on a foundation you can already trust.
Who you'd be working with
No account manager, no handoff to a junior team. The person you scope the work with is the person who builds it.
Founder & Principal Engineer
I'm a senior data warehouse developer. My background is the deep end of enterprise data work — dimensional modeling, large-scale synchronization between systems, and the kind of pipelines that quietly run a business every night without anyone thinking about them.
Data Supernova Inc. is where I bring that discipline to AI work. The industry is full of impressive prototypes standing on data foundations that won't hold. I build the foundation, then the thing on top of it.
Contact
A short description of the problem is enough to start. If it isn't a fit, I'll say so and point you somewhere better.
[email protected]