NotebookLM-inspired AI Document Platform
Academic team project combining FastAPI, PostgreSQL, Docker, AWS and generative AI for document processing and summarization.
Overview
A containerized application inspired by NotebookLM, with a React frontend, FastAPI backend, PostgreSQL persistence and generative AI integration.
Problem
Create a workflow for extracting and summarizing information from uploaded documents while keeping responses grounded in the available material.
Constraints
- Coordinate document processing, persistence and AI requests across application components.
- Keep local and cloud environments reproducible through containers and versioned configuration.
Approach
Combined a FastAPI backend, PostgreSQL data model and LangChain-based document workflow in a containerized application architecture deployed on AWS.
Key Decisions
Use FastAPI as the backend framework.
Its typed request models and asynchronous support fit the API and AI integration workflow.
Use Docker for environment parity.
A shared container configuration kept local development and deployment environments consistent across the team.
The initial setup added configuration work, but reduced environment-specific differences later.
Tech Stack
- Python
- FastAPI
- LangChain
- PostgreSQL
- Docker
- AWS
- React
Result & Impact
Delivered a working academic prototype connecting document ingestion, relational persistence and AI-assisted summarization in a reproducible containerized environment.
Learnings
- Grounding generative AI responses in uploaded document context.
- Coordinating relational persistence with an AI processing workflow.
- Containerized deployment and environment parity.