I'm building Oddit, a developer credibility platform. The idea is that the best proof of what an engineer can do is the code they've already written, not their resume or their GitHub stars. So I built an AI audit engine that actually reads your code and judges it the way a senior engineer would.

You submit a repo and it scores your work 0–100 across four things: depth of engineering, how hard the problems were, implementation quality, and how much the work actually mattered. Every finding points to a real file and line, so it's not just a number you have to trust.

It also goes past "how good is this code" into "what kind of engineer is this person." It reads your commit history to surface work patterns, like how you break down problems, how disciplined your iterations are, whether you self-correct, whether you actually finish what you start. Eleven different signals, each backed by real commits, plus the notable judgment calls behind your work.

All of that turns into a public portfolio you can share with anyone. And hiring teams use the same engine from the other side to search and vet candidates by what they've actually built.

It's live: DM to try