Lakmal Liyanage

13 years ago I started in QA doing what everyone did: running test cases, filing bugs, signing off releases. It worked, mostly. But the model had a ceiling — quality scaled with headcount, and headcount never scaled fast enough.

The shift happened gradually, then all at once. Manual regression gave way to automation. Automation gave way to shared platforms. The question stopped being "did QA test this?" and became "does the system make it hard to ship broken code?" That reframe changed everything about how I work.

Today I lead quality engineering orgs — SDETs building coverage, and platform teams building the tooling that lets developers own quality without a QA bottleneck at the end. Current focus: AI-enabled quality tooling. LLMs for test generation, smarter failure triage, coverage expansion without linear headcount growth.

This blog is where I write about that work. What worked, what didn't, and the decisions behind both.