At Early-Stage AI Startup (Stealth), I founded a research-driven AI initiative focused on LLM reasoning, evaluation, and prompt engineering. I lead research workflows from experimentation through prototype development and deployment, and deliver applied AI prototypes for early-stage clients.
In confidential AI research projects, I built interfaces and evaluation frameworks for assessing AI-generated outputs, and analyzed LLM reasoning with structured prompts and verifiable reward metrics. At Kantar, I developed APIs and applications with Angular and .NET, refactoring legacy modules to reduce load times by 25% and integration issues by 41%.

