Pitch candidates with proof, not just a CV
Good candidates get filtered on paper. Vettr gives you a 60-minute debugging session that shows how they actually think—evidence you can attach to your pitch.
Product demo coming soon
A report that shows what they actually did
Not pass/fail. Not a score. A timeline of how they investigated, what they tried, what they found—compared to what an LLM does with the same problem.
Candidate vs. LLM Baseline
| Behavior | Candidate | LLM Baseline |
|---|---|---|
| Investigated before changing code Used logs, traces, or metrics to understand the problem | ✓ Yes Opened Grafana within 15 min | ✗ No Jumped straight to code changes |
| Noticed the data volume anomaly Nova returns 5000+ features vs ~50 expected | ✓ Yes "Nova gets called 1600 some times" | ✗ No Never compared clients or volumes |
| Found the LIKE pattern bug LIKE '%Nova%' also matches NovaPay | ✗ No Saw the symptom but didn't trace it back | ✓ Yes Pattern-matched without understanding why |
| Applied a concurrency fix Parallel or batched database calls | ✓ Yes Used asyncio.gather | ✓ Yes Same approach |
Bottom line: The candidate demonstrated genuine investigative instincts—starting with observability, noticing anomalies the LLM missed entirely. Not LLM-assisted pattern matching.
For recruiters
Attach it to your pitch. Get candidates past the CV filter.
For hiring managers
See how they debug before you commit to an interview loop.
How it works
Real problem, 60 minutes
Production-style bug. Real tools, including LLMs.
We capture everything
Code, queries, investigation path. No screen share.
You get the report
Review in 5 minutes, not 60.
See what a Vettr report looks like
Drop your email. We'll send you an example report from a real session.