I Built DostStudy, a Private AI Study Buddy for a Classmate

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built

I built DostStudy, a private AI study buddy designed around a common
problem faced by a college student: difficult textbook explanations can
take too long to understand.

DostStudy turns a study question into a simpler explanation and can also
summarize material or generate a quick quiz.

The interface supports English, Hindi, and Hinglish so the learner can
choose the explanation style that feels most natural.

Demo

GitHub repository:
https://github.com/godGamerz812/doststudy

Demo video:
https://youtube.com/shorts/WfbHxrnq8AE?si=ITySnc1B8-Ssphwq

Code

Source code:
https://github.com/godGamerz812/doststudy

How I Built It

DostStudy uses:

• Gemma 3 as the open-weight AI model
• Ollama for model serving
• Python for application logic
• Streamlit for the user interface

The application is built around the AI model rather than using AI as an
optional add-on.

Why Does Open Innovation Matter?

Open innovation matters because the AI model is not locked behind one
closed provider.

DostStudy can use an open-weight model through Ollama, which gives the
project more control over the model and how it is run.

For a study application, local inference can also be valuable because
study material does not have to be sent to a third-party AI API when the
model is running locally.

The model can be swapped, the prompts can be changed, and the application
can be adapted without rebuilding the entire product around a single
closed API.

That flexibility is the main reason open AI was important to this project.

My Agent Session

Prize Categories

• Best Use of Gemma

Thanks for participating!