AI HR screening system
70% less screening time
The problem
Screening a stack of resumes against a job description took a recruiter most of a day, and the results changed with who was reading.
What I built
An n8n workflow watches the inbox, extracts each resume, scores it against the role with an LLM and writes a ranked shortlist with reasons to a sheet. This is a demo rebuild with sample CVs.
Key decisions
Structured output only
The model returns a fixed schema, so bad output is caught and retried instead of reaching the sheet.
Human stays in the loop
The workflow ranks and explains. A recruiter makes every final call.
Fallbacks on every call
Failed model calls retry, then route to a manual review queue.
Results
70%
less time spent screening
Need something like this?
