GENSEARCH

Inspiration

I wanted a smarter way to search through large PDFs and get accurate answers quickly using AI.

What it does

GENSEARCH lets users upload PDFs and ask questions. It finds the right info and gives smart answers using OpenAI.

How we built it

Built with Python, LlamaIndex for indexing, and OpenAI GPT for answering. Simple UI for uploads and queries.

Challenges we ran into

Handling large files, improving search accuracy, and integrating LlamaIndex with OpenAI smoothly.

Accomplishments that we're proud of

Working end-to-end RAG system, accurate answers, and a clean, user-friendly interface.

What we learned

RAG pipeline, prompt tuning, semantic search, and LLM integration.

What's next for GENSEARCH

Support more file types, add chat history, deploy online, and improve accuracy.

Built With

• faiss
• javascript-frameworks/libraries:-llamaindex
• langchain-apis:-openai-gpt-api-ui-platform:-streamlit-other-tools:-pymupdf-(pdf-parsing)
• languages:-python