PathOS

Inspiration Getting a job is hard. But knowing what job you actually want is harder. Up to 80 percent of college students change their major at least once. They waste thousands of dollars and add years to their graduation time. We looked at existing career tools. They all focus on the end of the journey. They help you format a resume or apply for an internship. But what if you are applying for the wrong thing? We realized the root problem is a lack of self-discovery. We built PathOS to solve this.

What it does PathOS is not a boring career survey. It is an AI-powered identity discovery game.

Dynamic Scenarios: The app drops you into real-world and academic dilemmas.

The Decision Mirror: As you make choices, our AI analyzes your behavior in real-time. It tells you how you think.

The Future Simulator: After the game, it does not just spit out a major. It gives you a side-by-side comparison of your best career paths.

Micro-Tests: It gives you a 7-day action plan with small, safe tasks to test the career out in the real world.

How we built it We wanted this to feel like a premium tech product, not a school project.

We built the frontend using Next.js and TypeScript.

We used Tailwind CSS for a dark-mode, professional design.

We used Framer Motion to add fluid swipe animations, making it feel like a game.

The core engine is powered by the Google Gemini API. We wrote strict system prompts to make the AI generate custom scenarios and output them in a pure JSON format.

Challenges we ran into Moving from hardcoded mock data to a live AI was our biggest challenge. We had to ensure the Gemini API always returned perfect JSON data. If the AI hallucinated the format, our React state would crash. We spent hours refining the prompt engineering to make the AI act like a strict game master. We also had to manage the loading states so the user never felt like they were just waiting for an API call.

Accomplishments that we're proud of We are proud that we did not build just another resume wrapper. We built a truly dynamic AI application. No two users will ever get the exact same sequence of questions. We are also very proud of the user interface. It looks clean, modern, and highly professional.

What we learned We learned how to integrate large language models directly into application state. We learned how to enforce JSON schemas on AI outputs. We also learned a lot about career psychology. We had to study psychometric frameworks to understand how to actually measure a user's interests accurately.

What's next for PathOS We want to integrate the YouTube Data API to automatically embed "Day in the Life" videos for the predicted careers. Our ultimate goal is to pitch this to our university's freshman orientation program to help undecided students find their true path on day one.

Built With

• framer-motion
• google-gemini-api
• groq
• next.js
• tailwind.css
• typescript
• vercel