A precision-first machine learning project that predicts whether a recipe is likely to generate high traffic. Combines data quality work, exploratory analysis, model comparison, and business-driven threshold optimization to achieve over 95% precision for homepage recipe selection.

Full-stack multi-vendor e-commerce platform with admin panel, vendor dashboard, order management, and commission tracking. Production-hardened with Redis caching (85%+ hit ratio), JWT auth, and load tested at 450+ RPS sustained throughput.

An end-to-end token-driven, privacy-preserving workflow system that controls the entire employee survey lifecycle from invite generation to response mapping, instead of just sending emails.

Production-grade CLO (Collateralized Loan Obligation) color data processing system for traders and analysts. Features automated scheduling, rules engine, Oracle/S3 integration, and an Angular dashboard — built as a client service project.

A stateful, cost-aware data ingestion system built on top of the BatchData API using Google Apps Script — engineered to solve the operational problems behind real-world data pipelines, not just fetch data.

Backend system for a tattoo studio booking and inbox management platform. Handles appointment scheduling, client communication, and studio workflow management through a RESTful API architecture.

Event-driven email automation system for an eCommerce business, solving the critical challenge of inbox deliverability — ensuring transactional and marketing emails bypass spam filters and reach customers reliably.

Beginner-friendly Jupyter Notebook demonstrating practical data import techniques in Python — reading flat files, CSV, and structured formats using built-in methods and pandas.

Exploratory data analysis on a real Los Angeles crime dataset. Covers data cleaning, statistical visualization, and extraction of actionable insights into crime trends and patterns.

Comprehensive analysis and forecasting project for silver prices using Yahoo Finance API historical data. Includes EDA, visualization, and a 3-month forward price prediction model.

ML-powered web app built with Streamlit to predict survival likelihood based on age, gender, and socio-economic status — demonstrating end-to-end ML model integration into a web interface.

A complete fraud detection case study focusing on business interpretability, cost-sensitive threshold tuning, and real-world deployment considerations — not just raw accuracy metrics.

Exploration of categorical data handling in Python — covering One-Hot, Label, and Ordinal encoding techniques, missing value strategies, memory optimization, and visualization best practices.

End-to-end ML project performing customer segmentation using KMeans clustering, with an interactive Streamlit dashboard for real-time segment analysis and business insight extraction.