Who Am I?
I'm an AI Quality and Evaluation engineer with 16 years in QA. I've built a QA function from scratch, led teams across three countries, and moved from manual testing to automation, then to security, and now to AI. Today I design evaluation strategy for Gemini-powered tools at Highspring, working with Google gTech, and I lead QA for the OWASP GenAI FinBot project.
What Do I Do?
I test AI systems the way real users and attackers will use them. I measure LLM output against ground truth, test agents for prompt injection and tool misuse, and build the automation that keeps them reliable.
How Do I Do It?
I work with product managers and developers from planning through launch. I don't trust a model until I've measured it: precision, recall, and how closely it agrees with human reviewers. I keep AI out of CI and check AI-written tests with mutation testing, because 100% coverage can still miss real bugs.
What Is This Blog About?
Practical lessons from building and breaking real AI systems, starting with FinBot. I write for QA engineers moving into AI, automation and security testing, and I share how to contribute to open source projects to build hands-on experience.