💡 Inspiration

Students often struggle with processing lengthy academic content and dense study materials. Preparing for examinations requires quick conceptual clarity and instant self-assessment tools. We built StudyPilot AI to solve this exact problem—breaking complex concepts down into bite-sized micro-lessons and generating practice quizzes on demand.


⚙️ What It Does

StudyPilot AI is an intelligent learning assistant designed to streamline self-paced education:

  • Micro-Lesson Generation: Breaks down technical topics into key bulleted points and practical analogies.
  • Instant Practice Quizzes: Automatically creates conceptual multiple-choice questions (MCQs) with instant evaluation.
  • Dual-Tab UI: Provides a clean Streamlit dashboard for seamless navigation between studying and testing.

🛠️ How We Built It

The application relies on a modular architecture leveraging Python and modern AI APIs:

  • Frontend: Built with Streamlit for a responsive web UI.
  • AI Engine: Integrated Google Gemini API (gemini-3.6-flash) via the google-genai SDK.
  • Environment Security: Secured sensitive credentials using python-dotenv.

Mathematical Formulation for Quiz Scoring

To evaluate student performance during dynamic quizzes, the system calculates score percentages using standard normalization:

$$Score (\%) = \left( \frac{\sum_{i=1}^{n} C_i}{N} \right) \times 100$$

Where:

  • $N$ = Total number of questions generated.
  • $C_i \in {0, 1}$ = Indicator variable where $C_i = 1$ if question $i$ is correct, and $0$ otherwise.

🚧 Challenges We Faced

  1. Prompt Engineering for Strict Output Formats: Ensuring the AI reliably outputs structured micro-lessons and clean MCQ options without unnecessary fluff required fine-tuning system prompts.
  2. API Key Security & Environment Setup: Managing sensitive API credentials safely using .env files while ensuring proper .gitignore configuration for public code sharing.
  3. Streamlit State Management: Maintaining state consistency across dynamic quiz generation and user responses without resetting UI inputs.

🎓 What We Learned

  • Practical integration of the new Google GenAI SDK using modern Python practices.
  • Designing human-centered AI prompts for structured educational workflows.
  • Streamlining rapid UI deployment and configuration for open-source submissions.

Built With

Share this project:

Updates

Submission history