DEV Community

Wahib EL KHADIRI
Wahib EL KHADIRI

Posted on

How a self-taught dev gets merged PRs into real open-source projects

I'm self-taught, 22, and I have merged pull requests in real open-source projects — a quantum-computing library, an AI coding-agent toolkit, security tooling, data stacks. No CS degree, no referral. Just a repeatable process. Here it is, no fluff.

Why bother (if you're job-hunting)

A merged PR is public, verifiable proof you can read a real codebase, follow its conventions, and ship. It beats a wall of tutorial projects on a resume, because a stranger with commit rights already said yes to your code.

The playbook

1. Find a genuinely small, unclaimed issue

Search GitHub for label:"good first issue" state:open in a language you know. Then filter hard:

  • 0 comments and no assignee → nobody's racing you.
  • The issue names the file and the fix → scope is bounded.

Skip anything vague ("improve testing", "add feature X"). Vague = subjective review = rejection.

# GitHub search that surfaces fresh, unclaimed issues
is:issue is:open label:"good first issue" no:assignee sort:created-desc
Enter fullscreen mode Exit fullscreen mode

2. Check for competing PRs before you start

Fast-moving repos get duplicate PRs. If someone opened one an hour ago, move on — I've lost that race twice and learned.

3. Reproduce, fix, and add a test

Clone, reproduce the bug, make the minimal change, then write a regression test. The test is what turns "trust me" into "here's proof." Run the full suite and note any failures that are pre-existing (verify with git stash).

4. Match the house style

Read CONTRIBUTING.md. Some repos require you to comment "I'll take this" first. Some enforce DCO (git commit -s). Some use Conventional Commits. Follow their rules exactly — reviewers notice.

5. Write the PR like a reviewer is busy

  • One sentence: what was broken.
  • The fix, and why this one.
  • Proof it works (test output).
  • Disclose AI assistance if you used it — honesty builds trust.

What I stopped doing

  • Chasing bounties. Most are dead (archived repos) or 200-person races. Terrible expected value.
  • Padding volume. Two clean PRs to reputable repos beat ten to 3-star toy projects. Quality is the signal.

The mindset

You're not "contributing to open source." You're solving one small, concrete problem for someone, cleanly, with a test. Do that ten times and you have a track record no bootcamp certificate can match.

Start today: find one 0-comment good first issue, and ship the smallest correct fix you can.

Top comments (0)