For developersYour repositories already know what you can do. Nobody has read them properly.
Pragmys analyzes your real projects to show the industry skills you actually have — CI/CD, containers, microservices, observability, not just the frameworks on your resume — and exactly which ones to build next.
Connect GitHub once. Get an evidence-backed, visual portfolio companies can trust, and a straight answer to the question you have been guessing at: why you did not get that interview.
Every row links back to the file, commit or workflow it was read from. Nothing here is self-reported.
From a repository to a claim someone can check
Three steps, no forms. Connect once and the rest reads straight from your repositories.
Connect GitHub
Your most active repositories are ranked and analyzed automatically. One consent screen, read-only, no separate connect step.
Signals become skills
Dockerfiles, workflows, migrations, queue configs and test suites map to a taxonomy of skills the market actually hires for — each scored on how consistently and recently it appears.
Act on the gap
Diff yourself against a real posting, see what is missing, and get the one project that closes the most gaps at once.
One scan in, a profile that holds up.
Six views, one evidence base
Skill graph, not a language chart
Every skill is a node sized by evidence and linked to the skills it shares repositories with. You can see your shape, not just your top language.
Cross-repo portfolio analytics
Build timeline, commit craft, project hygiene and your real stack — the patterns no single repository can show.
Role readiness against real postings
Paste a job description and get a requirement-by-requirement diff with the evidence behind each verdict, and the delta since you last checked.
A next move, ranked by leverage
Every gap comes with a concrete project — milestones, CI checks and all — not a vague suggestion to learn Kubernetes.
A public profile that holds up
One shareable link with every claim traceable to a commit. Recruiters can check it without taking your word for anything.
Three tools already claim to judge you. None of them read your code.
In the AI era every one of these is easier to game and less predictive than it was five years ago. Pragmys is the only one whose input is the work itself.
| Category | What it actually measures | Where it breaks |
|---|---|---|
| Screening platforms | How fast you solve timed algorithmic puzzles in a sandboxed editor under proctoring. | Nothing you do in production looks like this — and an LLM clears the take-home in a minute. |
| Profile builders | Surface GitHub metadata — stars, streaks, language bytes — formatted into a page. | Stars are popularity, not skill. It never opens a Dockerfile or a workflow file. |
| ATS & resume matchers Keyword parsers behind most job boards | Whether the words on your PDF overlap the words in the posting. | Rejects strong engineers over phrasing, and rewards whoever pasted the JD into their resume. |
| Pragmys reads the repositories themselves | What you have actually built and shipped — industry skills and craft signals read straight from your code. | Needs real repositories. If you have never built anything, there is nothing to read. |
Stop guessing why you were rejected.
One GitHub sign-in reads your top repositories, builds the profile and gives you a ranked list of what to build next. Only structure and names are uploaded — never your source code.
Sign in with GitHub →Free while in beta · read-only scope · delete your profile any time