Know who’s real before the interview.
Work Proof compares a developer candidate’s resume against their public work evidence (GitHub commits, shipped projects, live websites) and hands you a source-grounded interview pack. No guesswork. No gotchas.
Missing evidence is never treated as negative evidence.
The resume problem
Every shortlist is a guess.
You can’t tell from a PDF whether a candidate actually shipped the systems they describe. So hiring teams either trust the resume or burn hours verifying it by hand.
Resumes are marketing documents
Polished resumes are hard to assess without technical context.
AI detectors punish the wrong thing
Writing-style detectors flag good engineers for how they write, not what they built.
Manual deep-dives don't scale
Reading someone's GitHub properly takes hours. Impossible across 100 candidates.
How it works
From resume to interview-ready in four steps.
An automated pipeline audits each candidate’s public footprint and turns it into evidence you can act on, before a single interview hour is spent.
Drop in resumes
Bulk upload; the engine extracts jobs, dates, skills, projects, and links.
The engine audits the evidence
Pulls GitHub repos and commit history, visits websites, checks domain age, TLS, and liveness, verifies employer domains.
Claim-evidence matrix
Every resume claim gets a verdict: Partially supported or Could not verify, beside a deterministic score and plain-English flags.
Interview pack
Eight source-grounded questions with follow-ups, model answers, and "what to listen for", plus a versioned branded PDF report.
Evidence matrix
Every claim, checked against evidence.
The engine lines up what the resume says next to what the public record shows, and tells you plainly what held up and what didn’t.
| Claim | Evidence found | Verdict |
|---|---|---|
| Led microservices migration at Acme, 2023–2025 | 14 repos · 212 commits in period | Partially supported |
| Real-time dashboard, 10k daily users | No public artifact found | Could not verify |
| Mobile app, 50k downloads | Repo archived 2021 · no release tags | Could not verify |
Illustrative example. Gaps stay visible, especially when a developer’s work is private. A gap is a gap, never an accusation.
Features
A verification engine, not a vibes engine.
Every output is deterministic, versioned, and traceable back to the evidence it came from.
Bulk resume intake
100 resumes in, pipeline tracks queued → processing → complete with safe retry.
GitHub forensics
Repo inventory, commit history matched to candidate identity, immutable hashed evidence store.
Website & domain checks
Liveness, redirects, TLS, WHOIS domain age, framework fingerprinting.
Deterministic scoring
Score bands (strong / mostly / thin / unverified) plus JD role-fit scoring, frozen formula versions.
Interview Copilot
Eight-question packs with model answers; consent-first live interview sessions with audio transcription.
Reports that travel
Versioned web + branded PDF reports, hashed share links, expiring client packets, full audit log.
Who it’s for
Built for the people who sell and spend trust.
Staffing agencies
Resell verified, branded PDF reports to your clients as a premium service.
In-house technical recruiters
Screen developers on evidence, not keywords, before you spend interview hours.
Under the hood
Serious infrastructure, invisible to you.
A production-grade pipeline runs every check. You just see the verdicts.
FAQ