AgentMe

Inspiration

College students are overwhelmed and do not have the time or mental energy to constantly check job boards, scholarship sites, or campus portals. We wanted an AI version of the student that could do all of that searching quietly and intelligently in the background.

What it does

AgentMe creates an AI version of each student that searches for internships, scholarships, research positions, and campus opportunities based on their goals. It filters out noise, learns from each swipe, and delivers only the strongest matches in a simple swipeable mobile inbox.

How we built it

We built AgentMe as a mobile first React Native application using Expo with a smooth swipe interface inspired by modern consumer apps. The backend uses lightweight agents that simulate opportunity searches, extract key information, score relevance, and update preference models. The design features a calm color palette, smooth animations, and a personalized avatar generated from a student photo.

Challenges we ran into

The main challenge was designing an agent that feels smart while avoiding information overload. Keeping the inbox limited, relevant, and adaptive required careful filtering and preference learning. We also had to work around API limits and simulate safe and legal data collection.

Accomplishments that we are proud of

We created a polished mobile experience with a swipeable interface, a personalized agent avatar, and a curated opportunity inbox that feels genuinely useful. The app looks and behaves like a real product and the core architecture is ready for future automation.

What we learned

We learned how to design agent driven workflows that feel natural, how to build efficient mobile swipe interfaces, and how to guide AI behavior with structured prompting. We also learned how to balance simplicity, automation, and user control.

What is next for AgentMe

We plan to integrate real opportunity APIs, connect with systems such as Lehigh360, add notification digests, strengthen the agent learning loop, and eventually build a fully autonomous engine that scans the web for every student.

Built With

• chatgpt
• cursor
• gemini
• purplexity
• react.js
• typescript
• uxpilot