Data Science Portfolio

Machine learning

RAG Based Document Question Answering System using LlamaIndex

This project is a Generative AI-powered document search engine that enables users to ask natural language queries against a collection of unstructured documents (PDFs and DOCX). It leverages LlamaIndex, OpenAI LLMs, and HuggingFace Embeddings to perform semantic search over documents.

Simple Chatbot using OpenRouter

This is a self-project I built a simple Chatbot web app using streamlit and OpenRouter’s API. The goal was to create a lightweight, interactive chat interface powered by an LLM specifically, the Mistral Tiny model.

App Link

Telecome Churn Prediction

The primary goal of this project is to predict customer churn—whether a customer will leave the telecom provider (churn) or continue using the service. Predicting churn allows the company to retain customers by taking proactive steps to improve their satisfaction and engagement.

App Link

Bankruptcy Risk Prediction

This project contains information about various risk factors related to companies and aims to predict the likelihood of bankruptcy based on these factors. Each row represents a company with specific levels of risk in different areas, and the class column indicates whether the company is at risk of bankruptcy. The values in the dataset are categorized as follows:

1 represents a high risk.
0.5 represents a moderate risk.
0 represents a low risk.
App Link

Health Insurance Charges

This project contains information about individuals’ health, lifestyle, and demographic factors that could influence their health insurance charges. The project can be used to analyze and predict the cost of health insurance for an individual based on several key factors, such as age, sex, BMI, number of children, smoking status, and region.

App Link

Resume Classification

It is slightly simplified implementation of Kim’s Convolutional Neural Networks for Sentence Classification paper in Tensorflow.

Oil Price Prediction

The volatility and complexity of global oil markets make predicting oil prices a challenging yet crucial task for various stakeholders, including investors, policymakers, and industry professionals. In recent years, data science has emerged as a powerful tool to analyze historical trends, identify patterns, and build predictive models to forecast future oil prices.