DevMentor AI

Inspiration The inspiration for DevMentor AI came from a simple but powerful idea: what if every developer, from a student writing their first line of code to a professional tackling a complex problem, could have an AI-powered mentor by their side, providing instant feedback and guidance? The goal was to create a real-time learning platform that goes beyond simple syntax checking and offers deep, contextual insights, much like a human tutor would.

We envisioned a tool that could not only analyze code but also understand the developer's entire workflow—the screen, the voice, the diagrams—to provide a truly holistic learning experience. We were inspired by the advanced AI features in collaborative tools like Google Meet and wanted to bring that same level of intelligence to the world of software development education.

What I Learned This project was a deep dive into the world of multimodal AI and the practical challenges of building a responsive, real-time web application. The key takeaways were: •The Power of Multimodal AI: Integrating the Gemini API suite taught us that the future of AI is not just about text. By combining code analysis, video and audio processing (Gemini 1.5 Pro), and image understanding (Gemini Vision), we were able to build features that understand the full context of a developer's work. This project proved that we can create AI that sees, hears, and understands code, opening up incredible possibilities for education and productivity.

•The Importance of a Robust Backend: Building a service-oriented architecture with ASP.NET Core was crucial. Separating concerns into services (like NanoBananaService and GeminiService) made the code cleaner, more maintainable, and easier to debug. We learned the importance of dependency injection and how to properly register services in Program.cs.

•Frontend Complexity is a Real Challenge: As we added more features, the index.html file grew significantly. This led to several hard-to-debug issues, from the infamous "blank code editor" problem to API endpoint mismatches. We learned firsthand that careful organization, modular JavaScript, and meticulous debugging are essential for building complex single-page applications.

Built With

• api
• real-time
• which-allows-the-server-to-push-analysis-results-to-the-client-as-soon-as-they-are-ready.-2.-the-editor:-we-integrated-the-codemirror-editor-to-provide-a-rich
• with-the-gemini-api-at-its-core.-1.-foundation:-the-project-is-built-on-an-asp.net-core-backend-with-a-vanilla-javascript-frontend.-real-time-communication-for-instant-feedback-is-handled-by-signalr