← All Projects
// AI Tool
PDF Chatbot
Ask your documents anything — instant, grounded answers from any PDF.
Shipped — Personal AI Tool
📄
💬
🔍
// The Problem
Why PDF Chatbot
Long PDFs — contracts, research papers, manuals — are slow to search manually. Ctrl+F only finds exact words, not answers, and skimming a 50-page document to find one detail wastes time.
// The Solution
How PDF Chatbot Solves It
Upload any PDF and ask it questions in plain language. The app retrieves the most relevant passages and feeds them to an LLM, which answers based only on what's actually in the document — fast, accurate, and grounded in the source.
// How It Works
Inside the RAG Pipeline
01
PDF Upload & Parsing
The document is uploaded and its text extracted, page by page.
02
Chunking & Embedding
The text is split into chunks and converted into vector embeddings for semantic search.
03
Retrieval
When a question comes in, the most relevant chunks are retrieved by similarity search.
04
Grounded Answer
The retrieved passages are passed to the LLM API, which generates an answer grounded in the source text.
// Key Features
What's Inside
⚡
Real-Time Answers
Questions are answered in seconds, without waiting for a full re-read of the document.
🎯
Source-Grounded
Answers are built only from the retrieved passages — reducing hallucinated or made-up responses.
📄
Works With Any PDF
Contracts, research papers, manuals — any text-based PDF can be uploaded and queried.
🧠
Conversation Memory
Follow-up questions stay in context, so you can dig deeper without repeating yourself.
// Tech Stack
Built With
What I Learned
Building My First RAG Pipeline
This project was my first hands-on experience with retrieval-augmented generation — chunking strategy and retrieval quality turned out to matter more than the LLM itself. It's part of what pushed me to go deeper into machine learning during my Master's.
// Availability
Hiring, collaborating, or researching?
I'm always glad to talk about roles, freelance projects, or AI research collaborations.