Case study · Salil Lakra · 2026
Agile Turn Recruitment Suite
Applicant tracking for a real staffing firm — parse, match, hire.
Core contributor — 25 of 33 commits · 28 jobs · 17 candidates
Problem
A staffing firm needs one system for jobs, candidates, pipeline, interviews, and email — plus AI that can read resumes and rank fit, not a pile of sheets.
What I built
An internal recruitment suite: dashboard KPIs, jobs, applicants, Kanban, reports, AI search, interview scheduling, and an admin CRM. 107 API routes and a companion FastAPI ai-service for embeddings and NLP.
Architecture
Next.js App Router and API routes, PostgreSQL with Prisma and pgvector, Redis and BullMQ workers, NextAuth, Brevo, and a Python FastAPI service for embeddings and resume NLP.
My role
25 of 33 commits, merged via PR into the upstream repo. I shipped Gemini parsing, Google OAuth, invite-only registration, RBAC scoping, bulk resume upload, Brevo email, SSE parse progress, and GCP Terraform. The initial platform and the CRM module were built by other contributors — this is not a solo claim.
Frames
