PathFinder AI

Inspiration

Many students and early professionals choose their career paths based largely on family expectations or limited awareness of available options. Often, people realize much later that there were alternative career paths better aligned with their interests, strengths, and personality. This lack of early, personalized guidance can lead to dissatisfaction and missed potential. PathFinder AI was inspired by the idea of empowering individuals with data-driven and personalized career insights, so they can make informed decisions earlier and pursue paths that align with both success and happiness.

What it does

PathFinder AI is an AI-powered career guidance web application that analyzes a user’s profile—including interests, hobbies, academic background, achievements, and skills—to recommend suitable career paths.

For each recommended career option, the platform provides: Relevant job roles and opportunities A structured preparation roadmap Curated online and offline learning resources

Guidance services required to achieve the selected career path The goal is to move beyond generic advice and provide actionable, personalized career guidance.

How we built it

We followed a structured, phase-wise development approach to stay focused and deliver a working product within the hackathon timeline.

Phase-wise execution: Phase 1: Initial project setup and environment configuration Phase 2: Backend development and database integration Phase 3: Integrating and testing mock responses using the Google Gemini API Phase 4: Frontend UI development and API integration Phase 5: Dockerization of the application Phase 6: Deployment and final testing

Tech Stack: Backend: Python (FastAPI) Database: PostgreSQL Frontend: React.js

We began by completing the backend logic, including API endpoints and LLM integration, followed by building the frontend interface and connecting it with the backend services.

Challenges we ran into

Managing scope creep: As development progressed, we continuously came up with new feature ideas. To stay on track, we consciously parked these ideas for future phases instead of overloading the MVP. Dependency and compatibility issues: We faced challenges with incompatible Python library versions, which required careful debugging and dependency management. First-time integration with LLM APIs: Designing prompts and handling structured responses from the LLM required multiple iterations and testing.

Accomplishments that we're proud of

This being our first hackathon, we are proud to have: Designed and delivered a complete end-to-end working application Successfully integrated an LLM-based career recommendation system Deployed a full-stack application with backend, database, and frontend working together Maintained a structured development plan and met our core goals within the given time

What we learned

Through this hackathon, we gained valuable hands-on experience in: Designing scalable backend APIs using FastAPI Integrating Large Language Models effectively for real-world use cases Managing databases and environment variables in deployed applications Breaking down a large idea into achievable milestones Making trade-offs between feature ambition and delivery timelines Collaborating, debugging, and learning rapidly under time constraints Most importantly, we learned how to transform an idea into a functional product within a limited timeframe.

What's next for PathFinder AI

In future iterations, we plan to: Provide region-specific learning resources, opportunities, and career scope Add detailed skill-gap analysis and personalized learning timelines Introduce continuous profile updates so recommendations evolve with the user Expand career insights with local job market trends and demand analysis Our long-term vision is to make PathFinder AI a living career companion that guides users throughout their professional journey.