Quantifying the cost of being wrong.
I build the quantitative infrastructure that turns complex data into decisions teams can act on. End-to-end systems that give you the confidence to defend your reasoning in the boardroom.
The NHS has missed its cancer target every year since 2015. I built a quantitative system that diagnoses why.
Isolating nationwide pressures from local execution. Mapping the truth step by step, so you know exactly which failures are structural and which are operational.
NHS Cancer Pathway Recovery: Bayesian Inference for Operational Triage
A live interactive system that maps why the 62-day standard has been missed — separating systemic capacity issues from trust-level performance, with full posterior uncertainty quantified at every level. Explore it below.
Two types of work. Same standard of delivery.
Built for environments where a wrong call is immediately costly — healthcare, financial markets, complex operations.
Where the analysis is the competitive advantage.
For complex, high-stakes decisions where guessing is catastrophic and the wrong answer is expensive.
When you just need things built properly, fast.
For teams that need clean, bulletproof execution — without a six-month consulting engagement. Excel cleanup, scraping, automation, dashboards.
Donovan Robertson
I overhauled and delivered the analytical engine for a high-pressure trading operation — orchestrating centralised data, statistical models, and executive dashboards under extreme time pressure. Now I deploy that same institutional standard to build the quantitative edge for organisations where decisions are expensive and the data is complex.
Every engagement is owned end-to-end: from raw data to live system.
Read my professional background →
Every system ships with these non-negotiables.
The standard, regardless of project size.
Decision-Grade
Calibrated probabilities, proper uncertainty quantification, honest about what the model knows and doesn’t know.
Production-Hardened
Automated pipelines, drift detection, reproducible builds. Designed for the long term — not a one-off analysis.
Full Ownership
End-to-end delivery from raw data to live system. No hand-offs to separate teams, no black boxes, no loose ends.
Insights from Donovan Robertson.
Have a problem where being wrong is expensive?
Let’s begin solving your data problem. Brief me on what’s the issue below.
Request my expertise