Dev.to WebDev šŸ›  Dev šŸ‘ 0 šŸ“– 1 min read

I built a tool that scores your GitHub work 0–100 without reading your code

We're at a tech event in Berlin this weekend with one challenge: help as many developers as possible get hired. šŸš€ So we built BuilderHunt. You connect GitHub and get a free assessment of your public projects: šŸ“Š a 0–10

We're at a tech event in Berlin this weekend with one challenge: help as many developers as possible get hired. šŸš€

So we built BuilderHunt. You connect GitHub and get a free assessment of your public projects:

  • šŸ“Š a 0–100 score across six checks: ships, builds real things, architecture, stack depth, consistency, engineering habits
  • šŸ’Ŗ your 3 to 5 strongest selling points, plus your gaps
  • šŸŽÆ the German startup roles that fit you, from about 140 real listings, with links to the original postings

How it works

It looks at your public repositories from the last 24 months, but only file listings, dependencies, config files and READMEs. No source code is read, and GitHub access is read-only. The score is computed from facts about your repositories. It is not a model guessing whether your code is good.

Built with

Next.js (App Router), Postgres on Neon, Drizzle, Auth.js, deployed on Vercel. Scans run from a Postgres task table instead of a queue. Two LLMs: a cheap one for the project scan, a stronger one for the highlights.

Where it stands

Early MVP, built in a weekend. 60+ developers have signed up today. No companies are on the platform yet, so nobody gets introduced to an employer today. That is the next step.

What I'd like from you

Try it and tell me where your score is wrong. Is a single number useful, or off-putting?

šŸ‘‰ https://app.builderhunt.app/?ref=devto

šŸ“° Read the original article on Dev.to WebDev

Originally published by Dev.to WebDev. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.