Web App · AI
An AI recruiting platform, built solo
An AI-powered applicant tracking system built for a tech recruiting firm: candidate sourcing, job-fit scoring, automated triage, and AI-drafted outreach. A 6-phase pipeline keeps AI costs in check.
01 · Overview
The Challenge
An independent tech recruiting firm was running candidate sourcing, scoring, and outreach across a set of disconnected tools. Triage was heavily manual, and because every step leaned on AI APIs, the cost of running the pipeline was unpredictable.
02 · Approach
What I Did
Designed and built an AI-powered applicant tracking system, solo. The workflow runs as a 6-phase pipeline that separates no-cost tasks from paid AI operations, so API spend stays controlled instead of accruing on every step.
Six or more serverless edge functions handle the core work: candidate sourcing, job-fit scoring, automated triage, AI-drafted outreach, and presentation generation. Contact-email extraction from GitHub and Hacker News profiles is automated as part of sourcing.
Six or more serverless edge functions handle the core work: candidate sourcing, job-fit scoring, automated triage, AI-drafted outreach, and presentation generation. Contact-email extraction from GitHub and Hacker News profiles is automated as part of sourcing.
6-phase
Cost-controlled pipeline
6+
Serverless edge functions
Daily
In production use
03 · Result
Outcome
In daily production use, running the recruiting firm's candidate pipeline.
Built with
ReactTypeScriptSupabaseVercelOpenAI APIAnthropic API