This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend
What I Built
I built StudyMate, an AI-powered study companion for my college friend who studies Computer Science/Engineering and has trouble studying from long notes and preparing for exams.
The idea was simple: take long technical study notes and turn them into easier-to-use revision material.
StudyMate has four main features:
📝 Summarize — turns long notes into concise revision notes.
💡 Explain — explains difficult concepts in simpler language.
❓ Quiz Me — generates practice questions for exam preparation.
🃏 Flashcards — creates question-and-answer flashcards for revision.
I wanted to build something small and practical that could help my friend spend less time going through long notes and more time actually revising.
Demo
🌐 Live Demo: https://studymate-bhrreds3idz6bs3nqkrcga.streamlit.app/
🎥 Demo Video: https://drive.google.com/file/d/1S9J_Cp9t6Xkvg9uYXDQCVnh2_01c1L2a/view?usp=drive_link
The demo video shows all four StudyMate features working.
Code
💻 GitHub Repository: https://github.com/codelift338-dev/StudyMate
The repository contains the source code, README, and requirements needed to run the project.
How I Built It
StudyMate is built using:
Python
Streamlit
Hugging Face Inference
google/gemma-3-4b-it
The core flow is:
Study Notes
↓
StudyMate
↓
Task-specific prompt
↓
Gemma 3 4B
↓
Summary / Explanation / Quiz / Flashcards
Each feature uses a task-specific prompt designed for that particular study activity.
For example, the summarization feature asks the model to identify important concepts, definitions, and exam-relevant information. The quiz feature generates practice questions, while the flashcard feature converts important concepts into question-and-answer pairs.
I built and tested the application on a Dell Inspiron laptop with limited hardware. Instead of trying to run a modern language model locally, I used Hugging Face inference so that I could build the application around Gemma without requiring a powerful GPU.
Why Does Open Innovation Matter?
The core intelligence of StudyMate is based on an open-weight AI model rather than being designed specifically around a proprietary closed model.
Using an open-weight model gives the project flexibility. The model can potentially be swapped or experimented with as the application evolves without redesigning the entire application around one proprietary AI provider.
For this project, open AI also helped with the hardware limitations of my laptop. I could build an application around Gemma and use Hugging Face inference instead of requiring expensive local AI hardware.
The current version is not an offline application, because AI inference is performed through Hugging Face.
However, using an open-weight model gives me the opportunity to experiment with different models and deployment approaches in the future, including local inference with a suitable smaller model.
My Agent Session
DevRelay was used during the development setup, but I currently don't have a saved DevRelay agent session to embed.
The project source code and development work are available in the GitHub repository.
Prize Categories
Best Use of Gemma
StudyMate uses google/gemma-3-4b-it as the core AI model powering its summarization, explanation, quiz, and flashcard features.
Top comments (0)