JobOps Copilot

Human-in-the-loop AI agent platform for job-search operations

JobOps Copilot

_Deployed project May 2026 - Present_

Built and deployed a human-in-the-loop AI job-operations platform that combines a CRM, job discovery, RAG, multi-step agents, and workflow automation. The system drafts and recommends, but does not auto-apply or send outreach without user approval.

Key Contributions

  • Built a Next.js 16 and React 19 frontend with an Express/TypeScript API and Python/FastAPI agent service
  • Developed LangGraph multi-agent workflows for job tracking, company research, interview preparation, skill-gap analysis, and outreach drafting
  • Added pgvector-backed RAG over resumes and job descriptions, multi-provider model routing, tool calling, and persisted agent outputs
  • Deployed the web and API services on Azure App Service with Azure PostgreSQL and the agent service on Azure Container Apps
  • Improved ranking correlation from 0.716 to 0.821
  • Added n8n, Make, and Zapier workflow integrations, evaluation gates, PII redaction, rate limits, cost controls, and prompt-injection defenses

Technical Stack

Next.js, React, TypeScript, Express, Python, FastAPI, LangChain, PostgreSQL, pgvector, Azure, Docker, n8n