Projects

RAG based Research Assistant

A Research Assistant Tool that utilizes a Retrieval-Augmented Generation (RAG) approach to provide precise and contextually relevant answers to queries based on a dataset of scientific papers. By combining advanced retrieval techniques and large language models (LLMs), this system delivers accurate and concise responses for research purposes.

Key Features:

1. Semantic Search: Implemented using FAISS (Facebook AI Similarity Search) for efficient and scalable similarity-based retrieval.
2. LLM Integration: Uses cutting-edge language models for response generation.
3. Fine-tuning Capabilities: Enhances performance by fine-tuning LLMs with a domain-specific query-response dataset.

Text Summarization using Deep Learning

A text summarization tool that uses deep learning techniques to summarize scientific papers. The tool is capable of automatically generating concise and informative summaries by extracting key information from lengthy texts efficiently, leveraging deep learning algorithms, thus aiding researchers in identifying relevant literature.

Key Features:

1. BART auto-encoder used for text summarization.
2. Model trained on scientific paper datasets to generate accurate summaries.

Food, Nutrients and Obesity rate

In this project, we analyzed the relationship between the consumption of specific dairy products like butter, cheese, cream, yogurt, and ice cream, and obesity rates across various age groups in 2017-2018.

Methodology:

1. Data Collection - Clean and process the original data .csv files via Python.
2. SQL Database Initialization - Execute SQL commands via Python to create the necessary tables and attributes according to the data model, and load the data into SQL Database.
3. Export for Analysis - Query the tables for analysis and export the results.
4. Visualization - Upload the exported files to Tableau and visualize the findings.