CareerPilot
Human-controlled AI job agent for college students that tailors verified resumes, tracks deadlines, and runs mock interviews.
Built at AgentX Hackathon 2026
Team Runtime: Sunny kumar Sinha, Parmeet Singh Saluja, Sachin Kumar Dev, MD Rehan Khan
Built with Python, FastAPI, LangGraph, React, TypeScript, Tailwind CSS, SQLAlchemy, PostgreSQL, SQLite, PyMuPDF, ReportLab
About the project
The Problem: College students in India apply to dozens of roles under tight placement deadlines, often resorting to generic resumes or AI tools that fabricate skills and auto-apply without oversight.
Our Solution: CareerPilot is a calm, human-in-the-loop job application workbench built for students. Users upload their PDF resume once to establish a verified profile—a strict source of truth.
How It Works: Students filter 500+ job listings with computed skill-overlap scores and AI-generated fit bullets. Clicking "Prepare application" triggers a 5-step LangGraph orchestrator that recalls past interview weak spots, reorders matching projects, compiles a clean tailored PDF resume via ReportLab, drafts a cover note, and generates an .ics calendar reminder. A built-in code validator strips any unverified skills, projects, or employers before saving. Crucially, the agent cannot submit applications—nothing moves to "Applied" until the student reviews the side-by-side diff and clicks "Approve and send." CareerPilot also includes persistent mock interviews that score answers (1–5) and remember weak areas across sessions.
Impact: Eliminates resume hallucination and accidental submissions while giving students a transparent, auditable log of every AI action.