Community Star

I'm a Senior Quality Engineer with 11+ years across QA and software engineering, focused on risk-driven delivery in complex systems.

I direct effort where failure matters: customer experience, system behaviour, and cost of change. I combine exploratory testing for high-risk behaviour and unknown failure modes with automation and test architecture for reliable checking.

Quality is an emergent property, best understood through deliberate, incremental change. I focus on creating the conditions for safe experimentation in order to learn from real behaviour and iterate based on evidence. This means building testability and quality culture: reducing friction to testing, cutting noise in quality signals, and aligning around shared criteria.

My doctoral research in audio signal processing concentrated on separating overlapping sources from a single recording, which is an underdetermined problem with many valid outputs and no single correct one. AI evaluation often has a similar shape: many plausible outputs, where the challenge is identifying the signals that matter.

I'm building public work in AI systems evaluation, starting with an evaluation of an LLM-based customer review analysis system for faithfulness and prompt injection robustness. It uses calibrated LLM-as-judge scoring for faithfulness and metamorphic checks for robustness, where an adaptive baseline means the check only fails when the injection changes the output.