oracleodds AI

OracleOdds AI didn't begin as a hackathon project it began as an experiment.

Months before the Build with Gemini XPRIZE competition, I built a small MVP to answer one question:

"Could statistical models help everyday sports bettors make better decisions?"

The prototype proved there was demand, but it wasn't an AI-native business. It relied on traditional prediction logic and lacked the autonomous systems, production AI workflows, and commercial infrastructure needed to scale.

When the Gemini XPRIZE hackathon launched, I made a deliberate decision not to submit the old project.

Instead, I treated the MVP as product research.

Starting during the hackathon, I rebuilt Oracle around Gemini, designing an AI-native architecture where Google Gemini became a core production component rather than an add-on.

The earlier MVP served only as validation that users wanted the problem solved. The business that exists today was rebuilt around:

Gemini-powered reasoning autonomous prediction pipelines AI-generated analysis production inference infrastructure subscription billing real customer acquisition continuous AI operations

In other words, the MVP answered "Should this exist?"

The hackathon build answered "Can AI operate this business?"

That transformation turned an experimental prototype into a revenue-generating AI company.