ATUL PANDEY

// IIT Madras · Data Science · Machine Learning

↓ SCROLL TO EXPLORE ↓

About_Me

I am a Data Science student at IIT Madras with hands-on experience in data analytics, machine learning, ETL pipelines, and web development. I have worked across multiple internships where I handled real-world datasets, automated workflows, and delivered insights to support business decisions.

I enjoy transforming raw data into meaningful insights and building clean, developer-friendly applications using Python and modern web technologies. I believe in the power of data to unlock hidden patterns — much like seeing through the Matrix.

Skills

🧠

Machine Learning

LightGBM, scikit-learn, TF-IDF, Feature Engineering, Classification, NLP

📊

Data & Analytics

Python, SQL, ETL, Data Scraping, Predictive Modeling, Data Analysis

🌐

Web Development

Flask, HTML, CSS, Bootstrap, REST APIs, JavaScript

⚙️

Tools & DevOps

Git, GitHub, Excel, MS Office, VS Code, Jupyter

Projects

🧠 Comment Category Prediction

Machine Learning

Python · LightGBM · scikit-learn · TF-IDF · pandas

End-to-end ML pipeline for multi-class text classification on ~198K rows. Engineered features including log transforms, vote ratios, temporal features, and word/character TF-IDF n-grams. Achieved Macro F1 Score of 0.83187 using LightGBM, tackling severe class imbalance (55% vs 4%) with mixed text and numerical data.

🏆 F1: 0.83187 📁 198K Rows 🤖 LightGBM
⟩ View on GitHub

🏥 Hospital Management System

Full Stack

Flask · SQLAlchemy · Bootstrap

Developed a full-stack web application to digitize hospital operations, manage patient records, and centralize administrative workflows. Improved data accuracy and reduced manual errors using a database-driven design.

⟩ View on GitHub

📊 Operational Efficiency Optimization

Data Analytics

Python · Data Analysis · Forecasting

Analyzed sales, inventory, and cost data to identify reasons for stagnant profits. Discovered utilization gaps and seasonal trends, and proposed data-driven strategies for demand forecasting and cost optimization.

⟩ View on GitHub

Experience

Contact