#1 Product of the week · Launched October 20, 2026
AI-Atlas

AI-Atlas

Map your AI skills from your GitHub code

FreeAICodeAnalytics6,866 impressions#1 of its week20 comments

Maker links to DevHunt

Comments

>log in to comment
  • Aditya Goel· 27d ago

    Alex, scoring AI skills from the actual code rather than stars, with an evidence link behind every score, is the part that makes it believable to a hiring manager. Are developers using it mostly for job applications, or to benchmark themselves? Search note: I googled "prove ai engineering skills from github projects" (Google, US settings). Page 1 is KDnuggets, a LinkedIn post, a GitHub topic page, two skill repos, a Medium post, a jsdev.space ranking by stars and a dev.to guide. AI-Atlas isn't on it yet, and going by the titles, none of them scores your own repos, so a page built for that search has room. I run AI Knows Us, so this kind of search check is what I do.

    • Pop Alexandru[maker]· 27d ago

      Thanks Aditya, that's exactly the bet: a score is only worth something if you can click it and see the code behind it. Honest answer, since we launched just over a week ago: most people use it to benchmark themselves. They map their GitHub, see where they land on the field leaderboards, and rescan after they improve a project. A smaller group is starting to use it for jobs, with the Atlas link on a CV or LinkedIn, and that's the use I want to grow, because the evidence links are built for a hiring manager. Really useful search note, thank you. "Prove AI engineering skills from GitHub projects" is pretty much what AI-Atlas is, so I'll build a page aimed at that exact search. Good to know page 1 has no tool that scores your own repos.

  • kris gu· 27d ago

    Scoring based on actual code rather than star count is a very good approach, especially since each score is accompanied by evidence links, which can be verified when translated. I'm curious how the training logic part in the six domains extracts evidence from the code?

  • Cristina Alexandrescu· 1mo ago

    Bravo Alex! I'm clenching my fists

  • Daniela Lung· 1mo ago

    Good luck Alex!

  • AndronacheCarm3n· 1mo ago

    Interesting idea. Congratulations Alex!

  • Jarad Peckenpaugh· 1mo ago

    Very cool concept! The Custom Project Roadmaps they offer are great to get an outside opinion on your Github projects. I have been using them as a sort of "feature suggestion" for some of my autonomous AI agentic projects.

    • Pop Alexandru[maker]· 1mo ago

      Jarad, thank you, man! 🙏 You've been with AI-Atlas since the very first week, and a lot of what's live today exists because you actually used it and told me what felt off. Using the roadmaps as feature suggestions for your agent projects is exactly what I hoped someone would do with them. Seeing you rescan overnight and climb to 72.1 made my morning. You're officially ahead of me on the leaderboard now, so I guess I have some building to do.. :) Really glad we crossed paths through this. Keep the feedback coming!

  • Iulia Popa Psiholog Cluj· 1mo ago

    Congrats` Alexandru! You did a really great job!

    • Pop Alexandru[maker]· 1mo ago

      Thank you so much, Iulia! 🙏 Your support means a lot, especially today. It's been a long few months building this, and kind words like yours make it worth it!

  • 71 guardvip· 1mo ago

    Interesting! Congratulations on this outstanding project! Your hard work, dedication, and attention to detail truly shine through. The result is remarkable and reflects professionalism, creativity, and passion. We will follow the project with great interest and look forward to seeing its progress and future achievements. Wishing you continued success!

  • Paulstil· 1mo ago

    I like it

    • Pop Alexandru[maker]· 29d ago

      Thank you, the support really means a lot🙏

  • Rus Vlad· 1mo ago

    I find AI-Atlas a really interesting project, especially because the leaderboard adds a competitive aspect that motivates developers to keep improving their projects. I also like the methodology behind the AI predictor, which makes the idea even more compelling. I’ll definitely keep an eye on how this project evolves!

    • Pop Alexandru[maker]· 29d ago

      Thank you Vlad, that's exactly the idea! The leaderboard is meant to push people to keep improving real projects, not just collect stars. Since launch I've added a board for projects too, so every project now gets its own page and rank in its field. If you map your GitHub, I'd love to hear what you think of your result. And we're launching on Product Hunt this Friday, I'd be grateful for your support there too :)

  • Alex· 1mo ago

    Interesting idea indeed, the fact that a leaderboard exists it makes it way more "competitive" and in some way needing you to always improve projects or always create new better ones, to stay on the top, with the predictor machine explained in the methodology. Really recommend to try it!

    • Pop Alexandru[maker]· 29d ago

      Thanks a lot, Alex! Staying on top really does mean building better things, and that's the point: the score only moves when the code does. Glad the predictor and the methodology page made sense, that part took the longest to get right. If you try it and something in your score feels off, tell me, every bit of feedback goes straight into the next update. Thanks for recommending it!

  • Luiz Cesar Spies· 1mo ago

    Nice Project, interesting the way it was built, II will keep accompanying the journey of this Project.

    • Pop Alexandru[maker]· 1mo ago

      Thank you, Luiz! You were one of the very first people to use AI-Atlas, and your early feedback shaped a lot of it. Having you follow the journey means a lot to me! 🙏

  • Pop Alexandru[maker]· 1mo ago

    Hi! I’m Alexandru, a 17 year-old high school student from Romania working toward becoming an AI researcher. I built AI-Atlas on my own because “AI engineer” is now on every profile, yet nothing shows whether the code behind it holds up. AI-Atlas reads your public GitHub and scores the code itself, not stars or followers. Before the AI judges anything, it checks what can be proven, like whether notebooks were actually run and whether tests can run at all. So, in about a minute you get your skills map across six AI fields, a global rank, your own shareable atlas, and a live card for your GitHub README.

AI-Atlas reads your public GitHub repositories and scores each AI project’s actual code, not stars, followers or commit counts. It maps your skills across six fields: Machine Learning, Computer Vision, Data Science, NLP & LLMs, AI Agents and MLOps. How it works: 1. Enter any GitHub username. No account is needed to map a profile. 2. AI-Atlas picks out the repositories that contain AI work, reads their code and checks the facts it can verify. 3. In a few minutes you get your Atlas. Claim it with GitHub to appear on the leaderboard. What you get: - An Atlas profile: a score for each field, every project rated on a 7-part rubric, and the evidence behind each score. - A live README card: one line of markdown that shows your Atlas on your GitHub profile and updates whenever you rescan. - A leaderboard: global and per-country rankings, for verified owners only. - A predictor: shows which improvement would raise each project’s score the most, so you know what to build next. Who it’s for: Students and developers who want to show what they’ve actually built in AI, and anyone who wants honest feedback on which of their projects are strongest and what to improve next. To keep scores honest, the scanner checks what it can prove before the AI judge reads anything: whether notebooks were actually run, whether tests import modules that exist, and what CI really executes. Those checks are given to the judge as facts. The judge only rates each project; a fixed, published formula turns those ratings into your score, so no README can talk its way up. AI-Atlas only reads public repositories and never runs your code. Claiming your Atlas uses a read-only GitHub sign-in. Free to map your profile. Built by a 17-year-old high school student in Romania.

More AI tools →80 free dev tools →

  • Map what you've built on GitHub
  • Creates concise AI skills profiles with evidence links
  • Analyzes architecture, implementation, and training logic

AI-Atlas scans public GitHub repos to score and map AI engineering skills across six domains.

for
Developers who want to showcase and benchmark their AI project experience.
pricing
freemium
works withGitHub

Key features

8 features of AI-Atlas
  • Score per AI field — Assigns a score for each of six AI fields based on the code you actually wrote.
  • Live README card — One line of markdown puts your Atlas on your GitHub profile, and it updates every time you rescan.
  • Evidence behind every score — Each project score links to the files and code that support it, so you can see exactly why it scored the way it did.
  • Global & local leaderboards — Verified owners are ranked worldwide and within their country, by the evidence in their code rather than followers or stars.
  • Improvement predictor — Shows which change would raise each project’s score the most, so you know what to work on next.
  • Roadmaps and what-if scenarios — Turn a project’s evidence into a prioritized improvement roadmap, or estimate how much one specific improvement would raise its score.
  • Verified facts before judging — Before the AI reads anything, the scanner checks what can be proven: whether notebooks were actually run, whether tests import real modules and what CI really executes.
  • Fixed scoring formula — The AI only rates each project on a 7-part rubric. A fixed, published formula turns those ratings into the score, so no README can talk its way up.

Use cases

  • Create a public portfolio that quantifies AI expertise for recruiters.
  • Find the biggest gaps in an AI project.
  • Compare your AI skills with developers worldwide.
  • Get a step-by-step roadmap to improve a specific AI project.
  • See which of your projects best show your AI skills before an internship, job or university application.

AI-Atlas pricing

  • Free$0Mapping, your Atlas with every score and its evidence, leaderboards, the README card and y
  • 5 credits$3 one-time5 improvement roadmaps or what-if scenarios. Credits never expire
  • 15 credits$7 one-time15 roadmaps or scenarios ($0.47 per credit)
  • 40 credits$15 one-time40 roadmaps or scenarios ($0.38 per credit)

AI-Atlas FAQ

Does AI-Atlas run my code?+

No, repositories are read as text only; code is never executed.

Do I need a GitHub account?+

No account is required to map public repositories, but signing in lets you claim a profile and join the leaderboard.

What data does AI-Atlas use to score projects?+

Only public repositories the account owns: code, notebooks, READMEs and papers, plus checks such as whether notebooks were run and whether tests work. Stars and followers alone earn nothing.

How are improvement roadmaps generated?+

An AI model turns the project’s stored evaluation into prioritized, project-specific steps. Your first roadmap is free; after that, each one costs one planning credit.

Are there any subscription fees?+

No. Mapping, scores and leaderboards are free. Planning credits are a one-time, pay-per-use purchase and never expire.

Can I export the roadmap?+

Yes, saved results can be exported as PDF, JSON, or Markdown, or opened as GitHub issues.

Edited by the makers, based on getaiatlas.com · Oct 12, 2026. Details may change; check the official site.

Around the web

Articles, videos and discussions

AI-Atlas in the news

Recent AI-Atlas news and updates from DevHunt Daily.

  • · DevHunt launch

    AI-Atlas maps developers' AI skills from public GitHub code

    AI-Atlas analyzes public GitHub repositories to score AI projects across six fields, providing a detailed skill profile and a leaderboard. It offers a markdown badge for GitHub profiles and suggestions for improving project scores. The service is free and runs without executing code, targeting students and developers seeking verifiable AI credentials.