AIProductEngineering

CANDI

Candidate Analysis & Discovery Intelligence

Jun – Aug 2025

Chrome Extension (GitHub)

Overview

CANDI started as an internal recruiter workspace built during my CGP Group internship in Shanghai, and grew into a two-part toolkit for smarter hiring.

The first part is a Glide + Airtable app that compresses the most time-consuming recruiting tasks (requirements capture, candidate scoring, interview insights, and report generation) into a single workflow. The second part is a Chrome extension I built afterward that lets anyone paste a job description, open a LinkedIn profile, and instantly get an AI-generated match score with a plain-English explanation of fit and gaps.

The Problem

Recruiters were drowning in fragmented tools: requirements captured inconsistently, screening notes built manually in Sheets, data entered redundantly into the ATS, and status updates scattered across chats and emails.

Approach

  • Shadowed recruiters and mapped the end-to-end workflow: sourcing → screening → notes → HM report → ATS entry.
  • Built the CANDI Glide + Airtable MVP with AI prompt packs for criteria extraction, per-criterion scoring, and report generation.
  • Piloted with Singapore and Vietnam recruiting teams on live roles; iterated on UI/UX and outputs based on real usage.
  • Designed and built a Chrome extension that reads the active LinkedIn profile, accepts a job description as text input, and returns a structured match score with reasoning.

Outcomes

  • Screening time per candidate reduced from ~25–30 min to ~12–15 min.
  • Hiring-manager report prep dropped from ~45–60 min to ~10–15 min.
  • Chrome extension enables real-time candidate evaluation directly on LinkedIn, with no context switching.
  • Single source of truth for requirements across the recruiting team.

Stack & Tools

Glide, Airtable, Glue, Claude, Chrome Extensions.

← Back to Projects