The latest articles on DEV Community by Shadow (@shadow16ua). https://dev.to/shadow16ua https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4155614%2F5fd14b76-d881-4419-976e-ec2cb40b8722.jpg https://dev.to/shadow16ua en Shadow Sun, 04 Oct 2026 14:39:41 +0000 https://dev.to/shadow16ua/studybuddy-ai-taming-gemma-2b-to-build-a-local-quiz-generator-for-a-friend-4mpo https://dev.to/shadow16ua/studybuddy-ai-taming-gemma-2b-to-build-a-local-quiz-generator-for-a-friend-4mpo <h2> Who I Built It For </h2> <p>My friend is currently preparing for university exams and technical interviews (focusing on databases and C#). Reading dry documentation gets boring quickly, so I wanted to build an interactive, multiple-choice quiz partner that tests their knowledge on any given topic. Since it's for a student, it had to be completely free and accessible without internet restrictions.</p> <h2> What I Built </h2> <p>I built <strong>StudyBuddy AI</strong> — a lightweight, completely local web application that generates 3-question multiple-choice quizzes on any subject. </p> <p>GitHub Repository: <a href="https://github.com/Shadow16Ua/StudyBuddy-AI" rel="noopener noreferrer">https://github.com/Shadow16Ua/StudyBuddy-AI</a></p> <p><strong>The Tech Stack:</strong></p> <ul> <li> <strong>AI Model:</strong> Google's <code>gemma2:2b</code> running locally via Ollama.</li> <li> <strong>Backend:</strong> .NET 10 Minimal API (C#) to handle prompting and JSON parsing.</li> <li> <strong>Frontend:</strong> Pure HTML, CSS, and Vanilla JavaScript (with a sleek dark mode).</li> </ul> <h3> 📸 Demo </h3> <p><a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F35sytt21dmaq4d3o89iw.PNG" class="article-body-image-wrapper"><img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F35sytt21dmaq4d3o89iw.PNG" alt=" " width="800" height="496"></a></p> <h2> Why Open Innovation Matters Here </h2> <p>Choosing an open-weight model like Gemma 2B over a closed API (like OpenAI) was crucial for this project for three main reasons:</p> <ol> <li> <strong>Zero Cost &amp; Privacy:</strong> My friend can generate hundreds of quizzes without worrying about API limits, subscription fees, or sending their study data to a third-party server.</li> <li> <strong>100% Offline Capability:</strong> It runs perfectly on a standard laptop CPU, meaning they can study during commutes or internet outages.</li> <li> <strong>The Engineering Challenge (Taming the 2B Model):</strong> This was the most interesting part. I quickly realized that small 2B models struggle to output complex JSON arrays (like 3 questions at once). It would constantly hallucinate structures or merge answers. Because I had full control over the local inference, I was able to completely redesign the backend pipeline. Instead of asking for 3 questions in one prompt, my C# backend asynchronously asks Gemma for <em>one</em> question, exactly 3 times in a loop, and then manually constructs a bulletproof JSON array. This guarantees perfect UI rendering every single time. I also shifted the "shuffle options" logic to the frontend to prevent the AI from confusing correct/incorrect indexes. </li> </ol> <p>Open source allowed me to iterate rapidly, observe the raw output locally, and engineer a robust wrapper around a lightweight model to make it perform like a much heavier one.</p> <p><em>Note: I'm submitting this project for the Best Use of Gemma category as well!</em></p> devchallenge weekendchallenge hf26challenge gemma