ByteBuddy

Inspiration As a student navigating a rigorous curriculum while managing severe ADHD, I often found myself overwhelmed. Trying to focus on scattered educational materials—textbooks, PDFs, lecture slides, and random YouTube videos—was a recipe for cognitive overload.

This struggle sparked the idea for ByteBuddy. We realized that this isn't just a neurodivergent struggle; it is a universal student struggle. We wanted to build a centralized intelligence that doesn't just "show" you content, but actively helps you grasp it, organizing the chaos of a curriculum into a clear, navigable path.

What it does ByteBuddy is an adaptive learning companion designed to streamline the study process. It transforms static educational content into a dynamic, personalized learning experience.

Personalized Roadmaps: Users input a topic or upload reference materials, and ByteBuddy generates a structured, step-by-step learning path.

AI-Powered Note Generation: It digests complex material and provides concise, structured notes that are easy to review.

Adaptive Quizzing: The platform generates interactive quizzes that evolve based on the user's performance. If you struggle with a concept, ByteBuddy identifies it and adapts future questions to reinforce that specific area.

"On-the-Go" Integration: recognizing that busy students aren't always at their desks, we built a notification system. Users can opt to receive bite-sized notes and revision quizzes via WhatsApp or Telegram, making spaced repetition easy and accessible.

How we built it Building ByteBuddy was an intensive full-stack engineering challenge.

Backend: We utilized Node.js and Express to handle our server-side logic and API routing.

AI Engine: We integrated Various open ai model to handle the natural language processing required for summarizing notes and generating context-aware quiz questions.

Database: We used Supabase to store user profiles, learning progress, and generated content.

Messaging Integration: To enable the cross-platform features, we utilized the N8N with gmail and telegram API to push content directly to users' phones.

Frontend: The user interface was built using to ensure a responsive and gamified experience.

Challenges we ran into Hallucination Mitigation: One of our biggest hurdles was prompting the AI model effectively. Initially, the model would occasionally "hallucinate" incorrect facts or generate irrelevant quiz questions. We spent significant time refining our system prompts and implementing verification layers to ensure concise and factual results.

Gamification Logic: creating a quiz interface that felt "fun" rather than like a chore was difficult. We had to balance the difficulty curve so users felt challenged but not discouraged.

Asynchronous Messaging: configuring the webhooks for the messaging platforms (WhatsApp/Telegram) to handle real-time user responses required complex state management on the backend.