Launching today
AI-Atlas evaluates public GitHub code—not stars or followers—to reveal what developers actually build. Every AI project gets an explainable score using a 7-part rubric, mapped across six fields: ML, Computer Vision, Data Science, NLP and LLMs, AI Agents and MLOps. Map any GitHub profile for free. Claim yours to explore file-level evidence, join opt-in global, country and project leaderboards, add a live README badge and get a free improvement roadmap. Built solo by a 17-year-old student.
Free Options
Launch Team / Built With
AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted
Hi Product Hunt! I'm Alex, a 17-year-old high-school student from Romania and the solo builder of AI-Atlas.
I started it because I kept asking myself a simple question about my own GitHub: do my AI projects show real skill, or just a few imported libraries? Stars and followers don't answer that. The code does.
So AI-Atlas reads the code. Enter any GitHub username and it:
finds the AI/ML projects among the public repos
scores each one on a 7-part rubric (technical depth, difficulty, originality, completeness, usefulness, educational value, reproducibility)
shows the exact files behind every score, so nothing is a black box
maps your work across 6 fields and builds your Atlas
Mapping is free and needs no account. If you claim your Atlas with GitHub sign-in (read-only, public data only), you get:
a place on the global, per-country and per-field leaderboards
your projects ranked on their own project boards
a live README badge
your first improvement roadmap for free: concrete, file-level steps to level up a project
Some honest context: it's in beta and only sees public repos an account owns. A score measures the public evidence, not the person. It launched on DevHunt this week and so far 86+ developers have mapped their work. Many of the features came straight from early users' feedback, like ranking projects, not just people.
I'd love your help in two ways:
Map your GitHub and tell me if the score feels fair. Disagreement is the most useful feedback I get.
Tell me what would make this useful for you, whether you're hiring, learning or showing your work.
I'll be here all day answering every comment. Thank you for checking it out!