I'm a software engineer working mainly in Python and Go. My experience spans backend software, API integrations, SQL-backed systems and connected hardware, from implementation through testing, commissioning and support.
My merged Go contributions address download reliability in Ollama, streamed-output performance in Crush and cross-platform CI in OpenTelemetry Go. I also build local AI tools and work with JavaScript, TypeScript and React. Inspect the contributions and validation.
Ollama / Charm / Crush / OpenTelemetry Go
Portfolio / Open source / LinkedIn / Email
Project details, contribution evidence and background
|
A native Mac workspace for local models, agents, workflows and project tools. Includes a bundled Python engine and local Ollama inference. SwiftUI · Python · Local inference |
A native app for remaining limits, reset times and local usage trends, with a scriptable CLI. API-equivalent cost estimates are kept separate from subscription billing. Swift · macOS · CLI |
For a frontend example, browser-os is a small React desktop simulation with window management and a virtual filesystem stored in localStorage. It is a frontend experiment.
Practical improvements to reliability, performance and testing. These are merged contributions to projects maintained by other teams.
| Project | Contribution | Evidence |
|---|---|---|
| Ollama | Detect stalled model downloads before the first response-body byte. | Merged #17259 |
| Charm / Crush | Remove quadratic shell-output rendering and filter language servers before searching PATH. | Merged #3381 · #3370 |
| OpenTelemetry Go | Add cross-platform compile checks to CI. | Merged #8634 |
| W&B RAI Toolkit | Test the OpenAI-compatible model adapter. | Merged #48 |
| Internet Archive / Open Library | Fix template scanning in worktrees without submodule metadata. | Merged #13533 |
| ORAS Go | Clarify repository constructor and scope-helper migration documentation. | Merged #1378 · #1377 |
Under review: ZAP #7785, reducing false positives from public fallback responses in 403 bypass checks. Browse open contributions →
Pull-request status checked 21 October 2026. Benchmarks describe the workload recorded in each pull request. Compile checks are separate from running tests on the target platform.
Python and Go are my strongest languages, and I prefer backend work. I also work across web interfaces, APIs, data and embedded-connected systems. My independent work includes Wixal, a native app with a persistent Python engine and SQLite storage.
At IntelliDesign I developed application, internal-tool and embedded software in Python, Go, JavaScript and C/C++. At Traka / ASSA ABLOY I built Python and .NET API integrations and supported SQL-backed access-control systems through commissioning and APAC technical support. My PowerData contract combined CRM automation, authorised security testing and DISP compliance support, leading a team of three.
I'm based in Sydney, Australia, and open to relocating to Canberra. I'm keen to keep learning alongside experienced engineers and contribute to software that people depend on.
- Trace failures through interfaces, APIs, data and connected hardware. Reproduce the problem before changing the implementation.
- Keep pull requests focused, explain the trade-offs and retain useful regression tests, benchmarks and check output.
- Document configuration, recovery and handover so someone else can operate the system.
- Keep AI assistance explicit, review generated changes and verify the result.
Main languages: Python and Go. Other experience: JavaScript / TypeScript, React / Next.js, SQL, MongoDB, SQLite, C/C++, Swift and some Rust. Delivery: Linux, REST / SOAP, Git, pull requests and GitHub Actions.
Have a software or integration problem to work through? Get in touch or explore my experience.