Product
Instant Match, without the scramble.
Search and rank candidates with semantic similarity, layered scoring, and automated outreach. Your sourcing system becomes repeatable instead of heroic.
Job posts and candidate profiles can be embedded and indexed for vector similarity search.
A weighted ranker combines vector similarity and candidate scoring to produce top candidates fast.
AI-generated multi-variant outreach with cadence scheduling and event-based open/reply tracking.
P50/P95/P99 sourcing time, response rates, and by-skill breakdowns are exposed via APIs and UI dashboards.
Match quality, outreach cadence, and auditability matter. Instant Match is built to stay measurable and defensible.
Scores shown are illustrative. Production timing and accuracy targets require validation.
- Embed: Generate embeddings for candidates and job descriptions.
- Search: Vector search returns relevant candidates quickly.
- Rank: Similarity + candidate score produce a match score and breakdown.
- Reach: Outreach sequences run on a cadence; opens/replies feed analytics.
- Measure: Sourcing speed and response metrics show where to improve.
Semantic search, match ranking, outreach cadence processing, and engagement tracking exist in the repo. Achieving specific latency and pipeline-size metrics requires external validation and rollout.