dr kalid Asrat

I’m a physician-scientist and AI systems architect with 20+ years of clinical and public health leadership experience. I specialize in structured evaluation of large language model (LLM) outputs in healthcare, focusing on clinical reasoning validation, safety auditing, fact-checking against authoritative sources, and rubric-based model benchmarking.

In my work, I actively look for unsafe reasoning patterns, hallucinated medical claims, and incomplete patient guidance in AI-generated responses. I translate clinical-quality thinking into repeatable review methods that teams can use to measure performance, detect risk, and improve reliability.

As the Founder & AI Systems Architect of OmniHealthAI, I designed 100+ healthcare AI agents, including clinical reasoning assistants and evaluation tools. I developed standardized evaluation rubrics for diagnostic reasoning, triage decision-making, and patient communication—using authoritative evidence sources such as CDC, WHO, NICE, and peer-reviewed literature.