Selected work / Owned and partner product
A sports intelligence product built around the ticket
A web application that turns a sports-betting screenshot into structured ticket analysis, research, virtual practice, results tracking, and membership experiences.
Connected journey
The challenge
A screenshot contains imperfect data. A useful report needs to preserve what the ticket actually says, add only supported research, protect paid material, and keep later accounting and results consistent.
The case study documents product and engineering work across the customer experience, AI extraction, access rules, billing, community, and operating tools. It does not assign every historical contribution to one person.
What was built
A web application that turns a sports-betting screenshot into structured ticket analysis, research, virtual practice, results tracking, and membership experiences.
The important decision
The AI model is one component inside a stateful application. Raw ticket facts, calculations, cited research, recommendations, entitlement, and final results are kept as distinct responsibilities.
What the evidence supports
An October 1 verification record reports 201 unit tests, 54 integration tests, eight independent PostgreSQL concurrency checks, and a passing production build. Its full browser suite recorded 45 passes and two known baseline failures.
The reviewed material does not establish active-user growth, paid subscriber totals, revenue, or independently audited prediction performance. Historical release checks are dated evidence, not a claim of current production parity.