developer-roadmap (roadmap.sh)
kamran-ahmed‘s collection of interactive career roadmaps — 363.7k stars, 44.7k forks, which makes it the most-starred thing in this wiki and one of the most-starred repositories on GitHub of any kind. It describes itself as “Interactive roadmaps, guides and other educational content to help developers grow in their careers”, and it powers the site roadmap.sh. Routed here from Telegram 2026-08-05.
Worth stating plainly: this repository is one of the three artifacts the hub named when it spun this spoke out, and until today it had never been ingested. The spoke was founded on a cluster that cited roadmap.sh and then built its thesis without it.
What it is
80+ roadmaps, split two ways. Role-based: frontend, backend, full stack, DevOps, DevSecOps, AI engineer, data engineer, data analyst, product manager, engineering manager, QA, technical writer, DevRel, software architect. Skill-based: languages (JavaScript, TypeScript, Python, Go, Rust, Java), frameworks and tools (React, Vue, Angular, Node.js, Django), and specialty areas (system design, Kubernetes, Terraform, blockchain, cyber security).
Each node on a roadmap carries a description, links out to documentation and recommended courses, and
marks where a choice between alternative tools exists. Content lives as markdown at
roadmaps/<roadmap-slug>/content/<topic-slug>@<node-id>.md, and “merged changes are synced to the
website automatically” — so the roadmap graph is the source and the site is a build artifact.
It is the test case the spoke asked for, and the thesis survives it
The founding thesis says curricula are shaped by the assessment behind them, and it named its own falsifier: a curated-pointer curriculum with no assessment behind it. That is exactly what this is.
There is no exam behind a “become a DevOps engineer” roadmap — no standardized DevOps interview the way there is a standardized coding interview. And the pedagogy is pure curation: a graph of topics with pointers out to other people’s documentation and courses. It teaches nothing itself.
So the prediction the thesis makes here is the one it was going to fail on, and it doesn’t. Set against the other four:
| Source | Exam behind it | Pedagogy |
|---|---|---|
| coding-interview-university | yes | curated pointers |
| awesome-system-design-resources | yes | curated pointers + graded problems |
| ai-engineering-from-scratch | no | build from scratch |
| nirdiamant-prompt-engineering | no | build from scratch (notebooks) |
| developer-roadmap | no | curated pointers |
The correlation breaks. Two of the no-exam curricula build, one curates. What the exam predicted was scope — coding-interview-university excluding frontend and SQL because the interview does — and developer-roadmap’s scope is set by something else entirely: job titles. Its organizing unit is a role a company hires for, not a syllabus and not a body of knowledge. That is a third shaping force, and the thesis needs it.
What it does not have, which is the same thing nothing here has
300,000 developers is the reach figure the covering write-up quotes, next to 363.7k stars. No completion data, no outcome data, no effectiveness claim anywhere in the repository — and unlike the other four, this one does not even have a finish line to complete. A roadmap is a map. Nobody claims you arrive.
That makes it the spoke’s fifth T3 and the fifth source measuring intent to learn rather than learning. meta-curriculum holds the standing caveat and this does not move it. Tiered T3 on the ladder’s own terms: the maintainer’s own repository, with a commercial adjacency (roadmap.sh runs a paid tier and a newsletter the write-up puts at 10k+ subscribers).