• 

  Bite Logo

Project Story: Bite - The Pocket-Sized Knowledge Companion

Inspiration

Inspired by the sometimes excessive amount of time I spend in front of a screen, I realized that many of the times I reach for my phone, it's for something mindless and easy-- like TikTok, Instagram, or other infinite-scrolling social media platforms. It was this realization that made me think-- what if I spent this time mindlessly scrolling working towards a goal, like becoming a better programmer? Enter Bite--the infinite-scrolling short-form educational content app.

Learning Curve

The journey of building Bite was full of enriching experiences and novel learnings. Understanding the intricacies of diverse APIs such as OpenAI's GPT-3.5, Pinecone, Wikipedia, and HuggingFace's Sentence Transformers was intriguing. It offered an opportunity to explore the endless possibilities of content generation and recommendation. Valuable experience was gained in building an engaging, cross-platform frontend with React Native, and developing a potent backend with Flask and SQLAlchemy. Navigating the subtleties of user authentication using Clerk and deploying applications on AWS and Vercel demonstrated the importance of secure and scalable app development. It was fascinating to witness firsthand the power of machine learning and AI in creating a personalized and adaptive learning environment.

Construction Process

The Bite platform is an amalgamation of various technologies. React Native (with Webpack) was utilized for constructing a user-friendly, cross-platform frontend. Flask, a lightweight yet robust framework, was chosen for backend development. SQLAlchemy facilitated seamless database interactions. Diverse APIs such as OpenAI's GPT-3.5, Pinecone, Wikipedia, and HuggingFace's Sentence Transformers were harnessed for content generation and recommendation. User authentication was effectively handled by Clerk. To ensure scalability and reliability, the application was deployed on AWS (backend) and Vercel (frontend).

Encountered Challenges

Integrating multiple APIs to ensure seamless operation was indeed a significant hurdle. Fine-tuning the AI to generate meaningful, context-aware educational content was an exciting challenge. Ensuring the content was not only interactive but also personalized required diligent work. Clerk was instrumental in overcoming challenges related to user authentication and secure data handling. Deployment of the application for scalability was another obstacle, which was successfully handled using AWS for the backend and Vercel for the frontend.

Bite represents a transformative way to perceive educational content, making learning an engaging experience.

Sponsor Utilization

• OpenAI: The GPT-3.5 API was used for content generation with Wikipedia and as part of the recommendation engine.
• HuggingFace: It played a vital role in providing models for vectorizing data.
• Clerk: It was instrumental for user authentication.
• AWS: It was used for reliable backend deployment.
• Vercel: It was used for frontend deployment, ensuring the app's responsiveness and speed.
• Pinecone: It was crucial in content similarity search and served as part of the recommendation engine.

Built With

• clerk
• flask
• huggingface
• javascript
• openai
• pinecone
• python
• react
• react-native
• sqlalchemy
• vercel