A senior engineer at a mid-size fintech recently got screened out before a phone call because her GitHub profile showed zero public activity. Her actual contribution history was substantial: three years of internal platform work, two open-source security patches accepted upstream, and a Stack Overflow account with 4,200 reputation points in the cryptography tag. The recruiter never looked past the follower count.
This is not a story about a broken recruiter. It is a story about how hiring teams read public work wrong, and how engineers who understand the correct reading can turn their actual contributions into undeniable evidence.
What a Merged PR Actually Tells You
A merged pull request into a project you do not own is the densest signal available in a technical hiring process. Stars on a repository tell you about marketing. A merged PR tells you about judgment.
When you read a candidate's PR, you are not checking whether the code compiles. You are asking a specific sequence of questions. Did they scope the change tightly, or did they try to fix the universe in one commit? Did they write a description that explains the *why*, not just the *what*? Did they respond to reviewer feedback by defending their decisions with evidence, or by silently applying every suggestion like a contractor afraid of losing a contract? The diff is almost secondary.
Here is a real example of the kind of PR description that should move a hiring committee:
```markdown
Why this change
The current retry logic in ConnectionPool.acquire() uses a fixed 500ms backoff. Under sustained load (>200 concurrent requests), this creates a thundering-herd reissue spike that saturates the DB connection limit. Measured: P99 latency goes from 180ms to 4.2s during recovery windows.
What this does
Replaces fixed backoff with full jitter as described in the AWS architecture blog (https://aws.amazon.com/blogs/architecture/exponential-backoff-and-jitter/). Added a configurable maxbackoffms parameter (default: 2000ms) so operators can tune without a code change.
What this does NOT do
This does not address the root-cause connection leak in issue #847. That requires a separate refactor of the transaction scope. I can take that on separately if the maintainer wants. ```
This description shows measured thinking, knowledge of prior art, conservative scope, and explicit awareness of what was left out. A candidate who writes PR descriptions like this has already demonstrated the core skill you are trying to assess in a system design interview. You have the evidence. Use it.
Contrast that with a PR that changes three unrelated files, has a description that says "fixes bug", and shows a history of silently incorporating every review comment without a single explanatory reply. That pattern tells you the person knows how to submit code. It does not tell you they know how to think.
The review thread matters as much as the diff. A candidate who pushes back politely, explains their reasoning, and then concedes when the reviewer makes a genuinely better point is a staff engineer in the making. A candidate who either never pushes back or who turns every comment into a defensive argument is showing you something real about how they operate under pressure.
How to Read Stack Overflow and Forum Answers as Signal
Accepted answers on Stack Overflow carry a different kind of weight than PR contributions, but they are still underread by most hiring teams. Skills Tech Network surfaces this work as structured evidence precisely because so many teams ignore it.
The first thing to check is not the score. It is the answer date relative to the question date. An accepted answer posted in 2019 to a question asked in 2018 on a problem that is still cited in 2024 is worth far more than a thousand-point answer to a question that trended for one week.
The second thing to check is whether the answer explains tradeoffs. A useful answer to a technical question usually looks like this:
- Here is the direct solution
- Here is why it works
- Here is when you would NOT use it
- Here is the alternative if your situation is different
An answer that only covers the first bullet is a code paste. An answer that covers all four is consulting-grade communication, and it costs the candidate nothing to demonstrate except that they actually possess the depth.
The cryptography tag example from the opening is worth dwelling on. High reputation in a narrow, specialized tag is a stronger signal than moderate reputation spread across generic tags like javascript or python. Depth in a hard, narrow domain is exactly what principal-level roles require and exactly what is hard to fake. You cannot accumulate 4,200 reputation points in cryptography without actually understanding cryptographic primitives, common failure modes, and how to explain them clearly to developers who are confused and under pressure.
Vanity metrics to discard: total repository stars (reflects project visibility, not contributor quality), follower count (social proof, not technical proof), contribution graph streak (measures consistency of git pushes, including README typo fixes). None of these tell you whether the person can design a system under constraints.
Metrics worth your time: merged PRs on projects with active maintainers who review carefully, accepted answers in narrow technical domains with high question difficulty, design documents or RFCs posted publicly that show a decision process, and code review comments left on other people's PRs. That last one is invisible to most hirers but findable on GitHub. A candidate whose review comments consistently identify edge cases or ask clarifying questions about error handling is showing you exactly the behavior you want on your team.
One practical note on bias. Engineers from countries with less GitHub cultural penetration, engineers from enterprise backgrounds, and engineers from teams with strong confidentiality requirements will systematically underperform on public contribution metrics even when they are exceptional. Use public work to raise confidence, not to create a floor. If the evidence is strong, trust it. If the evidence is thin, look harder before concluding it is absent.
A well-reasoned merged PR into a competitive open-source project is worth more than a whiteboard session. It is primary source material, produced under real conditions, reviewed by peers who had no incentive to be kind, and permanently auditable. Hiring teams that treat it as decoration are leaving the most reliable signal on the table.
Build a proof-backed profile
Skills Tech Network ranks technical talent by verified, demonstrated capability, not just resumes. If you have public contribution history worth surfacing, or you want to build a portfolio that holds up to the kind of scrutiny this article describes, start here.
*Public work is the only hiring signal that cannot be coached into existence the week before an interview.*