-
-
InternPilot AI — System Architecture
Inspiration
Finding internships can be difficult for students because opportunities are spread across different websites. Students also spend a lot of time comparing opportunities and deciding which ones match their skills.
This inspired us to build InternPilot AI — an AI-powered internship assistant that helps students find, evaluate, rank, and track internship opportunities in one place.
What it does
InternPilot AI helps students throughout the internship discovery process.
- 🔎 Finds current internship opportunities using web search
- 🤖 Uses an AI agent powered by AWS Strands Agents SDK
- 📊 Evaluates opportunities based on skills, location, and experience
- 🏆 Ranks internships using match scores
- 💾 Saves internships for later
- 📌 Tracks application status such as Saved, Applied, Interview, Rejected, and Offer
- 💬 Supports natural-language internship searches
How we built it
We built InternPilot AI using Python and Flask for the backend and HTML, CSS, JavaScript, and Bootstrap for the frontend.
The AI agent is built using the AWS Strands Agents SDK. The agent uses a Tavily search tool to find current internship opportunities from the web.
Mistral AI is used for language understanding and internship evaluation. After search results are collected, our backend evaluates and ranks opportunities based on their relevance to the student's requirements.
The workflow is:
Student → Web UI → Flask Backend → AWS Strands Agent → Tavily Search → Evaluation & Ranking → Results
We also implemented internship tracking so students can save opportunities and update their application status.
Challenges we ran into
One of our biggest challenges was connecting an AI agent with real-world internship search while avoiding fabricated opportunities.
We also faced challenges while integrating AWS Strands Agents, configuring AWS credentials, improving search quality, evaluating internship relevance, and building a reliable ranking system.
Another challenge was keeping the interface simple enough for students while still providing useful AI-powered functionality.
Accomplishments that we're proud of
We are proud of building a complete working internship assistant rather than just a chatbot prototype.
InternPilot AI can search for real internship opportunities, evaluate them, rank them, and allow users to track their application progress.
We are also proud that we successfully integrated the AWS Strands Agents SDK into a practical student-focused application.
What we learned
Building InternPilot AI helped us understand how AI agents can go beyond generating text and actually perform useful tasks through tools.
We learned how to:
- Build an AI agent using the AWS Strands Agents SDK
- Connect an agent with external tools
- Use web search for current information
- Evaluate and rank opportunities
- Build a Flask application around an AI agent
- Deploy and test a complete working application
- Design an AI workflow focused on a real user problem
What's next for InternPilot AI
We want to make InternPilot AI more personalized and powerful.
Future improvements could include:
- Resume-based internship matching
- Personalized recommendations
- Email and application reminders
- Better ranking and matching algorithms
- Persistent database storage
- More advanced agent workflows
- Additional internship sources
Our long-term goal is to make internship discovery faster, more relevant, and less stressful for students.
Log in or sign up for Devpost to join the conversation.