StudyDos Project Story

Inspiration

StudyDos was born from observing that students often struggle to balance the need for assistance with the importance of developing critical thinking skills. Instead of spoon-feeding answers, we envisioned an AI assistant that guides learners to discover solutions on their own—promoting deep engagement and independent problem solving.

What It Does

StudyDos is a full-stack, multi-modal academic assistant that:

• Guides Assignments: Provides tailored hints for assignments and projects without revealing complete solutions.
• Plans Studies: Generates detailed study plans to help students organize coursework and exam preparation.
• Assists with Code: Supports code completion for programming tasks using cost-effective AI models.
• Handles Diverse Inputs: Accepts file uploads and supports image recognition to process various learning materials.
• Integrates Course Material: Leverages both stored course content (via MongoDB/GridFS) and static data to deliver context-aware responses.

How We Built It

We developed StudyDos using a modern, multi-technology approach:

• LLM API (FastAPI):
  Our Python backend uses FastAPI with asynchronous OpenAI API calls to generate study plans, assignment hints, and code completions. Course materials are stored in MongoDB using GridFS.

• Backend (Express.js):
  A Node.js/Express server manages user queries, stores messages in MongoDB, and communicates with the LLM API for dynamic responses.

• Frontend (React with TypeScript & Vite):
  The React-based user interface provides a chat-style experience along with additional features such as deadline management and study plan displays.

• Environment Management:
  We securely manage API keys and database URIs using environment variables with tools like python-dotenv and dotenv.

Challenges and Lessons Learned

• Asynchronous API Integration:
  Adapting to asynchronous programming with FastAPI and OpenAI’s API required significant refactoring and a steep learning curve.

• Balancing Cost and Performance:
  We tuned our AI models (using GPT‑3.5-turbo for chat and a cost-effective model for code completions) to keep quality high while managing costs.

• Multi-modal Data Handling:
  Integrating file uploads and simulating image recognition introduced design challenges that ultimately enriched the user experience.

• User Intent Detection:
  Developing a unified endpoint that accurately distinguishes between study plan requests, assignment hints, and code completion queries was critical.

• Cross-Technology Collaboration:
  Working with Python, Node.js, and React reinforced the value of agile methodologies and coordinated teamwork.

What's Next

• Enhanced AI Capabilities:
  Integrate adaptive learning algorithms and personalized recommendations.

• Robust Database Integration:
  Expand course material management for even more tailored responses.

• User Feedback Loop:
  Build analytics and feedback mechanisms to continuously refine the assistant.

• Broader Multi-modal Support:
  Improve image recognition and file processing to support a wider range of educational resources.

• Educational Partnerships:
  Pilot StudyDos in real classroom settings to gather insights and drive further enhancements.

Figma Prototype

For a visual overview of our UI design, please check out our Figma prototype:

StudyDos Figma Prototype

StudyDos embodies our commitment to empowering students with intelligent, ethically designed AI that supports learning without compromising their ability to think critically.

Built With

• express.js
• fastapi
• figma
• mongodb
• python
• react