Project Overview
Mobile app that transforms academic papers into accessible formats using machine learning, making knowledge accessible to all.
React Native
Python
AWS
LangChain
PyMuPDF
BART-LARGE-CNN
Key Features
- Building a React Native app that reformats academic PDFs for mobile reading, using PyMuPDF and pdf2image for content extraction.
- Leveraging LangChain with BART-LARGE-CNN on AWS Elastic Beanstalk to stream chunked summaries, cutting server load by 50%.
- Designed a minimalist, scroll-friendly UI and integrated async queuing and serverless caching to reduce latency on large (60+ page) uploads.
Technical Implementation
PaperWave addresses the challenge of reading complex academic papers on mobile devices by implementing a sophisticated content processing pipeline:
- Content Extraction: PyMuPDF and pdf2image extract text and images from academic PDFs
- Text Processing: LangChain with BART-LARGE-CNN model processes and summarizes content
- Mobile Optimization: React Native app provides a responsive, touch-friendly interface
- Cloud Infrastructure: AWS Elastic Beanstalk handles processing with async queuing and caching
Performance & Optimization
The app is designed for performance, with server load reduced by 50% through chunked summaries and streaming. The async queuing system and serverless caching ensure smooth performance even with large documents (60+ pages). The minimalist UI design prioritizes readability and ease of navigation on mobile devices.