Anton Turbinov

Spending my early career in Customer Success, where I analyzed customer conversations gave me a sharp eye for natural interaction. I learned firsthand that all systems automated or not will only succeed when they are built around the nuances of real-world user behavior. Today, I leverage my 15-year IT background to build AI-native quality assurance systems from scratch. I spend my time engineering end-to-end automation workflows—using Playwright and TypeScript. Lately, I've been building "LLM as Judge" multimodal frameworks—essentially creating the evaluation layer where advanced models review and critique the work of other autonomous agents generating web and social content. I am a huge fan of the idea to make AI testing pragmatic and thorough so development teams have a clear, measurable data on how their AI is performing.