At Sama AI and Instadeep, I formulate complex, domain-specific prompts to stress-test LLMs on medical, biological, and clinical reasoning. I evaluate and rank model responses for factual accuracy, logical coherence, tone, and adherence to multi-criterion rubrics.
I also audit outputs for hallucinations, reasoning breakdowns, and safety or policy edge cases, then provide actionable feedback to MLOps and engineering teams. My work includes executing complex guidelines in an asynchronous, self-directed environment.
Previously, at A2D3 / Applied AI Africa, I audited structured datasets and documented error taxonomies and evaluation criteria. I study Medical Sciences / Nursing at the University of Nairobi, with coursework in anatomy, pathology, biochemistry, and research methods.
