In my independent projects, I designed and ran clinical prompts and multi-turn vignettes to assess the medical accuracy and diagnostic safety of ChatGPT (GPT-4), Claude, and Google Gemini. I evaluated responses against clinical standards including PubMed, NICE, and WHO.
I also wrote evidence-based reference answers and corrections for ambiguous medical test cases and pharmacotherapy edge cases. This work helped me identify issues such as missed contraindications and unverified dosage recommendations.
At Thika Level 5 Hospital, I studied the prevalence, causes, and risk factors of surgical site infections after Cesarean section. I curated clinical records and analyzed the data with SPSS and Microsoft Excel, then presented findings and visualizations to supervising clinicians and peer panels.
At AnalystLab Africa, I worked with healthcare and operational datasets, building interactive Power BI dashboards and automating reporting workflows in Excel. Iām also a final-year MBChB candidate at Mount Kenya University, Thika.
