At BQE Software, I built a 737-case offline evaluation bank and an LLM-as-judge harness for CoreAI Advisor, making evaluation a required check on every pull request.
I also owned the GA programme's integration foundations, including integration-branch CI, migration tooling with checksummed rollbacks, and release checkpoints. I co-own the CoreAI Quality Strategy, covering system-level SLOs and a definition of done for AI pull requests.
I built CORE Analyst's navigate-and-hand-off feature end to end, and helped lift next-step recall from 87% to 94%. I also shipped a help-center content pipeline and features for project-completion and billing metrics.
Before moving to CoreAI Advisor, I ran QA for BQE Software's CORE Android and iOS apps. I wrote over a thousand test cases, reported 557 defects, and built Kotlin instrumented UI automation for critical flows.

