Four years of production Python and cloud engineering, now building agentic AI systems. Currently completing a postgraduate certificate in Applied AI Solutions Development at George Brown Polytechnic. I ship things, not just notebooks.
An agentic AI system that reads government API documentation, decides which of 40,000+ open Canadian datasets are actually relevant to a question, pulls the real data, and returns an analysis, not just a metadata description. Built as a Chrome extension talking to a Flask backend on AWS Lambda.
Production agentic AI system combining RAG-based retrieval, memory, and MCP-style tool calling to automate multi-step workflows across tabular and unstructured data.
Four-model federated learning pipeline (BERT, LSTM, Autoencoder, Seq2Seq) for privacy-preserving fraud detection on Canadian financial documents. No raw data leaves the client system.
Modular ML evaluation pipeline benchmarking Pearson Correlation, Chi-Square, RFE, L1 regularization, Random Forest importance, and LightGBM across real-world datasets.
Transfer learning with MobileNetV2 on the PlantVillage dataset, classifying leaf disease across 39 categories. Built to learn CNN fundamentals end to end.
Looking for a Winter 2027 co-op or full-time role. Canadian Permanent Resident, fully authorized to work in Canada. Based in Kitchener, open to relocating anywhere in Canada, or remote/hybrid.