HireMatch
A vacancy and resume become a reviewable match map that exposes evidence, gaps and scoring logic.
Context
A final percentage alone gives recruiters little confidence. A useful decision requires visible evidence for each requirement, explicit information gaps and focused interview questions.
How it works
Split AI processing into controlled stages: vacancy structuring, human review of requirements, resume fact extraction and matching. OpenAI returns data through strict Pydantic schemas, while separate deterministic logic calculates the final score.
What it enables
The public service gives recruiters a transparent basis for judgment rather than an automated verdict: evidence, mismatches, gaps and interview questions live in one coherent workflow.
Key engineering decisions
Controlled AI processing stages
Vacancy analysis, resume parsing, matching and summarization have separate roles. Structured data with schema validation passes between the stages.
Review requirements before matching
The user reviews requirements and their importance before the final analysis, allowing misinterpretations of the vacancy to be corrected before scoring.
Scoring is separate from text generation
Matching connects requirements to resume evidence; the server calculates the score using fixed weights. AI prepares explanations and questions but does not make hiring decisions.