Champion – 4th National Data Science Summit
AI Chatbot Category
Our team proudly secured the Champion Award at the 4th National Data Science Summit: AI in Education, held at Daffodil International University. Competing with top innovators, we showcased our AI-powered Rag Based Chatbot, impressing judges with its automation, accuracy, and real-world impact in modern recruitment.
This project is an AI-powered chatbot built with Retrieval-Augmented Generation (RAG) capabilities, designed to extract context-aware answers from both plain text and scanned PDF documents. By integrating meta-LLAMA 3.1 (8B, 4-bit quantized) and running it locally using Ollama, the solution demonstrates how powerful AI models can be made lightweight and accessible — without relying on cloud APIs. Awarded Champion in the AI Chatbot Category at the 4th National Data Science Summit, this project impressed judges with its ability to merge local LLM inference, OCR, and vector search into a seamless knowledge assistant.
Showcase
What it does
- Plain Text + Scanned PDF Support
- Offline Inference with Meta LLaMA 3.1
- Vector-Based Context Retrieval
- Conversational Agent with RAG
- Modular, Extensible Design
Goal
To create a fully local, low-cost intelligent chatbot that can read and respond using real-world documents — even scanned ones — enabling deployment in data-sensitive or offline environments like schools, clinics, or rural offices.
Tech stack
- LLM
- Meta LLaMA 3.1 (8B, quantized 4-bit)
- Runtime
- Ollama + LangChain
- Vector DB
- FAISS
- Document Handling
- PyMuPDF, EasyOCR
- Frontend
- Streamlit / Gradio
Team
Tasrif Nur Himel
Project Lead, LLM Integration, Vector DB, OCR
Mojahid
App Developer
Links & references
Have an AI problem worth solving?
Tell me what you are trying to build. I will tell you honestly whether AI is the right tool, what it would take, and how I would approach it.
Typical reply within 24 hours
Working from Dhaka, Bangladesh · remote worldwide