I'm still early in my career and keen to learn more. These are projects I built to practise, and each one says what it measured and where it falls short. Based in Melbourne.
Machine learning and computer vision PyTorch, TensorFlow, scikit-learn, ONNX · Computer vision, NLP, LLM agents, recommenders · CNNs, transfer learning, model evaluation
Software engineering Python, C/C++, TypeScript, JavaScript, Java, C# · FastAPI, Node.js, React Native, .NET · Embedded (STM32, ESP32), Docker, Terraform · Azure, GitHub Actions, Git, Scrum
Data analysis and visualisation SQL, PostgreSQL, SQL Server, dbt, pandas · Power BI, Tableau, Excel
| Repo | What it shows |
|---|---|
| llm-project | A text-to-SQL agent over 202,554 Victorian road crashes. 82% correct on a 22-question gold set (measured). |
| edge-vision-lab | A scene-split bird, drone and neither classifier. Plain int8 broke it; a partly int8 model recovered most of it (measured, laptop CPU, no board). |
| facial-expression-lab | Thesis approach re-tested on JAFFE. Accuracy fell from 0.864 to 0.573 once the test set held out whole people (measured). |
| mini-city-brain | 1,577,675 real traffic readings through an API, queue and warehouse. 15 of 15 data checks passed (measured, one laptop run). |
| asia-commodity-trade | Australia's share of Indonesia's wheat imports fell from 57.6% to 40.6% between 2015 and 2025 (UN Comtrade). |
| cs-notes-and-games | Free computer science study site and puzzle game: 104 levels, 68 chapters. Built with Claude Code, disclosed in the README. |


