My Projects

Challenge

Design a single system that could take raw sales data, clean and structure it, run useful statistical analysis, and present the results through both a dashboard and API.

Solution

Built an end-to-end Python pipeline with ETL preprocessing, star-schema storage in SQLite, ARIMA forecasting, RFM segmentation, anomaly detection, a Streamlit dashboard, and a Flask API for reusable access.

Challenge

Build a reusable pipeline that can collect job posts, normalize messy listing data, and score resumes against multiple skill dimensions in a practical way.

Solution

Created a modular 5-stage workflow for scraping, parsing, normalization, matching, and output generation, with independent stages that made new source integration much faster.

Challenge

Create a scheduling system that handles services, clients, calendars, and appointment requests while keeping both admin and client experiences clear.

Solution

Built separate admin and client flows with React and Tailwind, used Supabase for auth and data, and added a Node.js layer for secure AI-powered booking insights.

Challenge

Build a shopping-style UI with reusable structure and responsive layout.

Solution

Created reusable React components, product listing views, and deployed on Vercel.

Challenge

Run hand tracking in-browser with stable updates and readable UI output.

Solution

Used TensorFlow.js with smoothing logic and simple overlays for interaction feedback.

Challenge

Make algorithm flow easier to understand for students and beginners.

Solution

Built interactive controls and visual steps for algorithm simulation and comparison.