Senior software engineer based in Wellington, New Zealand.
I work across software systems, backend engineering, infrastructure, cloud, developer tooling, and production debugging. I am strongest when a problem is ambiguous, crosses multiple layers, or needs careful investigation before a fix can be trusted.
My current focus is AI-assisted systems engineering: using advanced models as an engineering harness for reproduction, root-cause analysis, implementation, adversarial testing, and verification.
My Workflow:
Issue discovery ↓ Issue scoring / prioritisation ↓ Repository acquisition + environment setup ↓ Codebase mapping ↓ Failure reproduction ↓ Instrumentation ↓ Hypothesis generation ↓ Parallel investigation agents ↓ Candidate patches ↓ Tests ↓ Benchmarks ↓ Static analysis ↓ Adversarial review agent ↓ Patch refinement ↓ PR preparation ↓ Maintainer feedback ingestion ↓ Revision ↓ Engineering case study
Upstream PR: netty/netty#17391
Investigated a pooled direct-memory regression affecting HTTPS workloads with slow consumers. Isolated the backpressure failure to SslHandler, implemented a bounded TLS wrapping mechanism, added a regression test, and validated the repair against the complete Netty handler module:
- 15,635 tests
- 0 failures
- 0 errors
- JDK TLS compatibility verified
- BoringSSL/OpenSSL compatibility verified
A model-independent engineering control plane for directing AI systems through:
reproduction → evidence → root cause → design → implementation → verification
The objective is not code generation. It is reliable AI-assisted systems engineering with persistent state, evidence gates and human approval.
Public case studies demonstrating production-style debugging, CI repair, environment diagnosis and repository rescue.