Pathwise 🔮

A developer built Pathwise, an AI-powered career decision platform, for a friend experiencing post-graduation career paralysis. The app uses open-weight models such as Gemma to simulate career outcomes, generating Risk Radars, five-year Project Roadmaps, and Confidence Scores in a Bento-grid dashboard, and was built and deployed over a single weekend using open-source tools.

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend https://dev.to/challenges/hacktoberfest-weekend-2026-10-01 Pathwise is an AI-powered career decision platform designed specifically for my friend who is currently struggling with severe post-graduation career paralysis. Instead of traditional, static career counseling that leaves them anxious, Pathwise acts as a "Future Me Simulator." It takes their skills, interests, and industry trends to simulate concrete outcomes for different career paths. It breaks down their long-term goals into visual Risk Radars , actionable 5-year Project Roadmaps , and Confidence Scores —all presented in a clean Bento-grid dashboard to turn abstract career anxiety into actionable clarity. Live Demo: https://pathwise-cl9l.onrender.com https://dev.toh { https://github.com/ursarthak001/PATHWISE https://github.com/ursarthak001/PATHWISE } I built the entire application using open-source tools and open-weight models: Open innovation was absolutely critical to this project because career aspirations and anxieties are highly personal . By building on open-weight models like Gemma , I wasn't locked into expensive, opaque proprietary APIs that might quietly ingest my friend's vulnerable data. Furthermore, the open-source ecosystem—from Python's robust libraries to the accessible Hugging Face hub—allowed me to build, iterate, and deploy a professional-grade AI application over a single weekend without spending a dime on enterprise software . Open innovation democratizes the ability to solve hyper-specific, personal problems using world-class intelligence. Throughout this time-crunched weekend, I heavily utilized my local AI agent to troubleshoot localhost connection errors, bypass terminal Git issues via GitHub Desktop, and successfully deploy the repository securely to the cloud before the deadline.