Work
A selection of full-stack applications, client websites, AI tools, and ML systems I've built.
Production-grade e-commerce platform — admin dashboard, product import pipeline, inventory management, Redis-backed task queues, and Docker deployment.
Live healthcare platform delivering supplements and prescription items across Canada — WooCommerce backbone with 50+ products, PHIPA compliance, EMR integration, and automated prescription workflows.
Production-grade RAG system for LLQP exam candidates — PDF ingestion, Qdrant vector search, OpenAI embeddings, grounded answers with source citations, web chat widget, and Telegram bot interface.
Revamped a Canadian pharmacy website — modernized infrastructure, built automated form workflows with instant clinic notifications, and set up a scalable fertility services subdomain.
Professional website for a long-standing Toronto optical clinic — fully responsive across all devices, promotional popups, and local SEO to drive foot traffic and bookings.
Site-wide redesign for a premium interiors brand — standardized header/footer, launched a new Services page, improved technical SEO, and added a custom 404.
Lightweight portfolio for a Toronto ceramic artist — real-time Etsy product syncing, embedded gallery view, and seamless click-through to purchase on Etsy.
Open-source Telegram bot that classifies music files by genre — Librosa feature extraction, TensorFlow neural network, and a conversational Telegram interface.
Tagged and analyzed 2,000+ podcast episodes by guest, topics, and user interest. Built a Telegram bot with 150+ active users/month and a React + Flask web interface.
Trained and compared five models (MLP to CNN) on the MNIST handwritten digit dataset — reached top 5% on Kaggle with batch normalization, data augmentation, and global average pooling.
Local Streamlit app that processes any podcast — YouTube URL or audio file — through transcription, chunking, embedding, summarization, emotion profiling, keyword extraction, topic clustering, and hybrid semantic + emotion recommendations.
Deterministic tile-based Pac-Man with five AI controllers (Random, Greedy, Ghost Avoidance, A*, Behavior Tree), a headless evaluation engine, and a Reports UI for multi-agent comparison over 1,000+ seeded episodes.
Suite of 6+ real-time computer vision modules — edge detection, object tracking, and video stabilization deployed on webcam feeds with <100ms latency.