Recruitment Operations Platform

Overview

A role-based system of record connecting business development, delivery managers, and recruiters around the same client and candidate data, replacing fragmented manual tracking. Built end-to-end: a Next.js front end, a .NET (Aspire) API, and a stored-procedure-first SQL Server layer on Azure Kubernetes.

Challenge

1. 01Model a long, multi-actor lifecycle (request, allocation, recruiter, candidate pipeline, close) with role-based access across 5+ roles.
2. 02Add AI value without blocking the request flow: requirement analysis, Boolean search strings, and résumé-driven screening questions.
3. 03Run reliably and affordably across test/UAT/prod on shared infrastructure, with per-environment config that could not drift.

Approach

1. 01Architected the system end-to-end: a Next.js 16 / React 19 front end, a .NET 10 Aspire-orchestrated API, and a Dapper, stored-procedure-first SQL Server layer.
2. 02Built a consistent service/controller pattern with Azure AD JWT auth, global exception handling, and fire-and-forget background LLM tasks.
3. 03Owned the delivery path: Azure DevOps CI/CD with build-once/deploy-everywhere, Key-Vault-sourced ConfigMaps, and a namespace-isolated AKS cluster behind nginx ingress with managed TLS.

Outcomes

1. 01Shipped the recruitment lifecycle plus a three-phase sales-monitoring module (pipeline, activity logging, reviews, dashboards).
2. 02Delivered 18+ API controllers and dozens of stored procedures with pagination, soft-delete and audit trails, and AI-assisted screening.
3. 03Stood up a repeatable multi-environment AKS deployment that became the template for the rest of the estate, and cut cluster cost by roughly half.