Placement Co-pilot
Autonomous AI agent that reads your resume, closes skill gaps, rewrites ATS bullets, and seeds real jobs into your pipeline — one run.
Built at AgentX Hackathon 2026
Team finding_natural_acceptance: Pratham Vithalani, Manan Hindocha, Sanchit Sawant
Built with Next.js 16, React 19, TypeScript, Tailwind CSS v4, LangGraph.js, LangChain Core, Google Gemini 2.0 Flash, xAI Grok, Groq Whisper, Supabase, MediaPipe, Piston API, GitHub REST API
About the project
Engineering students waste weeks guessing which skills to learn, manually tailoring resumes, and tracking applications in spreadsheets — with no feedback loop. Placement Co-pilot replaces that chaos with a real agentic pipeline.
Select your target role (Backend Engineer, AI/ML Engineer, etc.). Upload your PDF resume. A Gemini-powered autonomous agent runs a multi-step tool-calling loop: it parses your resume, scores your fit against the role, identifies missing skills, rewrites your project bullets to match ATS keywords, injects a personalized readiness study plan, and seeds live job openings into a Kanban tracker — all in one run, with human-in-the-loop approval before any resume changes are applied.
The agent uses a real agentic loop built on LangGraph.js and LangChain core, with Gemini 2.0 Flash as the reasoning engine. Tool outputs are fed back into Gemini's context window across multiple steps — making it a genuine agent, not a scripted chatbot. xAI Grok serves as a live fallback when Gemini is unavailable.
Additional features: mock AI interviewer with STAR method evaluation and GitHub deep-grill questions, a DSA coding arena with 30 curated problems powered by the Piston execute API, in-browser camera analysis via MediaPipe, and a Demo Mode that runs without any API keys or Supabase setup.
Built for final-year engineering students navigating placement season with zero recruiter guidance.