Recubix
Resume Matching

Find the right candidate,
not just the closest one.

Semantic matching that understands context, not just keywords. From 200 resumes to a ranked shortlist in under 3 minutes.

Try matching freeStart free

How it works

JD
Job description uploaded or AI-generated
→
Parse
Resumes extracted to structured profiles
→
Embed
OpenAI embeddings for semantic matching
→
Index
Profiles stored in Pinecone vector DB
→
Rank
70% criteria + 30% semantic score per candidate
→
Output
Top 100 ranked list with gap analysis

Scoring model: 70% weighted criteria match (skills, years, location) + 30% semantic embedding cosine similarity via Pinecone.

📂

Batch resume upload

Upload up to 500 PDFs or DOCXs at once. Recubix extracts skills, experience, contact details, and deduplicates — then ranks every candidate against your JD in under 3 minutes.

🔍

Global semantic search

Type in plain English: "Senior backend engineer in Bangalore with 5+ years Python and fintech experience". Get the top 100 candidates from your talent pool ranked by semantic + criteria fit.

✍️

AI job description generation

No JD? Describe the role in a sentence and Recubix generates a structured JD with must-haves, nice-to-haves, and skill tags — ready to use for matching in seconds.

Sample ranked output

RankCandidateMatch %Gap
#1

Priya S.

Senior ML Engineer · Bangalore

96%None
#2

James T.

ML Platform Lead · London

91%Fintech −
#3

Aiko M.

Backend Engineer · Singapore

84%Python exp −1yr

Illustrative example. Real output includes full skill breakdown, contact details, and resume link.

From 200 resumes to a shortlist in 3 minutes.

No spreadsheets. No manual filtering. Just the top candidates, ranked.

Start matching