- Problem
- Release risk gets decided by gut feel while the evidence sits unused in the repository.
- Role
- Architect, engineer, operator
- Stage
- Six sprints in; census and classification before production exit
- Stack shape
- Web + worker monorepo, authenticated with row-level security
Kinetik
Release-risk intelligence for software teams: turn the evidence a repository already holds into a risk read a human can act on before shipping.
Claim: most teams ship on instinct. Commits, pull requests, review threads, and change history already contain the signal — but almost no one turns that into a usable answer to "how risky is this release?"
What's built: a working system that ingests GitHub evidence, reasons over the change surface, scores release risk, and generates changelog and release-risk intelligence at the decision point. Authentication and row-level security are in place; the pipeline runs as a web + worker monorepo.
Receipt: six sprint cycles of working software in private alpha. Not screenshots of a prototype — an operating system-under-test with real repository data flowing through it.
Status honesty: per ASV Labs discipline, Kinetik is being censused into the project register and classified before its production exit continues. It's the front-runner in the workflow lane and the leading candidate for first product revenue — which is exactly why it doesn't get to skip the process. Launch claims come after the gate, not before.