EV battery performance analysis
Analyzed electric motorcycle battery swap and telemetry data to identify degradation patterns and forecast fleet usage.
I turn complex data into practical decisions. With 7+ years across data science, actuarial analysis, and business intelligence, I build predictive models, forecasts, analytical tools, and dashboards that help teams act with confidence.
I work across the full analytical lifecycle - from data preparation and statistical modelling to forecasting, automation, visualisation, and decision support. I also bring deep actuarial expertise, including pricing, product analysis, claims and underwriting, and parametric insurance.
More recently, I've applied this mix of technical and domain knowledge to AI evaluation and quality assurance: designing rubrics, verifiers, test cases and evaluation tasks, and assessing AI systems across LLM, computer-use, audio and video workflows.
Tools
My work spans Python, SQL, R, and Excel, from data preparation and model validation to reporting in Power BI, Tableau, and Looker. I work in GCP (including BigQuery), build deep learning models in TensorFlow and PyTorch, use Git for version control, and apply agent evaluation frameworks to test AI systems. Throughout, I focus on communicating findings clearly to non-technical teams.
Applied projects across machine learning, actuarial analysis, healthcare, mobility, operations, iGaming, and sports analysis. Each project links to its source repository, report, dashboard, or case study.
Analyzed electric motorcycle battery swap and telemetry data to identify degradation patterns and forecast fleet usage.
Compared Nairobi sub-counties on facility access, bed availability, and selected health services against population and target ratios.
Analyzed case-processing operations and agent performance, including review delays, resolution times, workload, reopens, and efficiency trends. The report turns these findings into recommendations for improving case handling and team performance.
An interactive Tableau dashboard presenting iGaming metrics through visual analysis.
An interactive Looker dashboard exploring ride-hailing data.
Built a claims development triangle, calculated loss development factors, and projected ultimate losses from payment data.
Created a web app that uses recency-weighted historical performance to estimate outcomes across leagues and teams.
Implemented the chain ladder reserving approach in R as a companion to the Python project.