Franklin Karimi

I'm a medical doctor and data scientist working in clinical AI and health data science, focused on predictive models, risk stratification and clinical decision support that can actually survive deployment in African health systems. I combine hands-on machine learning in Python with a clinician's understanding of why patients disengage, why adherence fails, and what a ward can realistically act on.

I work at the intersection of these three groups that rarely speak the same language: clinicians who know where care breaks down, data teams who can model it, and the people who decide what gets funded and deployed. I take clinical problems, frame them as data problems, build and evaluate the models in Python with scikit-learn, and carry them through to deployable services. Evaluation is where I put most of my weight. Calibration, threshold selection against the real cost of a missed case, subgroup performance, and an honest benchmark against what staff already do without a model. I would rather establish early that a model is not worth deploying than help scale one that has not earned its place. That judgement is the part that takes clinical years to build, and it is what I bring to a technical team.

As co-founder and clinical lead at Infarmatech, a chronic disease digital health venture, I own the clinical logic: risk stratification rules, referral criteria, encounter templates and dashboard indicators, plus bilingual low-literacy screening standards for the community-facing tool. I specify what the system does automatically and where a clinician must stay in the loop.

That work is grounded in clinical frontline practice. I am a practicing medical doctor in Kenya, caring for patients across emergency, medicine, paediatrics, obstetrics, surgery, psychiatry and ICU while building in clinical ML