Piyush Pandey

Building a full-stack journaling app with Spring Boot REST APIs and JWT authentication, using PostgreSQL with vector embeddings to enable AI-based semantic search over journal entries. Structuring the backend with a layered architecture (Controller–Service–Repository–Entity) and DTOs for clean API contracts and secure data handling.

Engineered time-series features (lag values, rolling averages, temporal indicators) from multi-station pollutant and weather data sourced via OpenAQ and Open-Meteo APIs to improve model predictive power. Trained and compared Random Forest and Linear Regression models for PM2.5 prediction, applying outlier removal and chronological train-test splitting, achieving R² = 0.84 and MAE of 5.4 µg/m³ on t

Developed a machine learning classification system to predict heart disease risk using patient health attributes, with data preprocessing, categorical encoding, feature scaling, and model evaluation across multiple ML algorithms. Deployed an interactive Streamlit web application using a trained K-Nearest Neighbors (KNN) model to provide real-time heart disease risk predictions from user-provided c

Grade: 77.5%

Pursuing a Bachelor of Technology in Computer Science & Engineering with a focus on Data Science.

Grade: 80.6%

Completed AISSCE (Class XII) with 80.6%.

Grade: 84%

Completed AISSE (Class X) with 84%.