How an established public health researcher with 150 papers tested whether academic rigor could translate to production ML systems — built models processing millions of data points, and discovered that building tech products was more energising than publishing journals.
Established Public Health Researcher – 150 papers, tenure
Accra, Ghana
Role: Lead Data Scientist / Principal Data Scientist
Duration: 2 terms across 6 months
Dr. Kwame had spent his career excelling in academia. 150 peer-reviewed papers, tenure at a respected university, citations that shaped public health policy. He was undeniably successful in the research world. But a question had been nagging him for years: what if his expertise could create impact at a different speed? Academic research moved in cycles of months and years. Tech products shipped in days. Could his analytical rigour translate to production machine learning systems? Could he enjoy building products as much as publishing papers? He needed to find out.
A global HealthTech company building a disease surveillance and outbreak prediction platform — used by WHO and public health agencies to detect emerging threats.
A ClimaTech company building a climate risk modelling platform for insurers — predicting regional climate risks to inform insurance pricing and coverage decisions.
Both companies had strong engineering teams building the product infrastructure. Dr. Kwame's role was to build the ML models that powered the predictions.
Two terms in different domains tested whether the translation from research to production was replicable — not a one-off success but a genuine capability.
"I'd excelled in academia — 150 papers, tenure. But could I excel building a scalable health surveillance platform? Two SaaS products tested me in ways journals never did. Now I'm building ML models that help predict outbreaks in real-time."
— Dr. Kwame, Accra
An early-stage HealthTech SaaS was building a pandemic preparedness platform and needed a founding data scientist — someone who could build the ML foundation from scratch. The HealthTech team from Dr. Kwame's first term recommended him. His track record of shipping production ML models, combined with his deep epidemiological expertise, made him the obvious choice. Dr. Kwame joined as founding data scientist, building the platform's prediction models from the ground up.
Dr. Kwame's analytical capability was immediately valuable in production ML. But he had to learn to ship fast, iterate based on user feedback, and accept "good enough" over "perfect."
Users of the platform provided immediate, actionable feedback that improved models faster than any academic review process could.
Success in HealthTech alone could have been luck. Success in ClimaTech proved it was a genuine capability — and made Dr. Kwame uniquely valuable.
This hypothetical scenario shows how academics and researchers can test whether building tech products creates the impact they're looking for. Your testing ground could start today.