The most important behaviour emerges at the boundaries.
Most of my work sits in the space between mathematics, data, and real-world decision-making. That usually means dealing with uncertainty, incomplete information, and systems that behave differently depending on their state.
Whether I’m looking at a financial market, a climate system, or a startup ecosystem, the structure of the problem is often surprisingly similar:
There is hidden state you cannot directly observe.
Small changes can have disproportionately large effects.
Stability is often temporary, not guaranteed.
I don’t think of my work in terms of isolated techniques.
I think in terms of systems - how they evolve, where they break, and what sits underneath the surface.
I work with systems that don’t behave nicely.
My niche lies where linear intuition fails and dynamics matter more than averages.
That includes:
Dynamical systems and chaos
Stochastic processes and noise-driven behaviour
Metastable systems and regime shifts
Critical transitions and tipping points
Nonlinear feedback structures
Operator-theoretic representations of dynamics
What this really means in practice:
I’m interested in what happens before a system changes state - not just after the change is obvious.
A lot of traditional modelling starts with equilibrium assumptions.
I’m more interested in what happens when equilibrium is no longer a useful concept.
This is the foundation of everything else.
I build frameworks that don’t assume stability.
How do you make robust decisions when the underlying system may change its rules?
Most real-world decisions are made under uncertainty.
Not uncertainty in the abstract sense - but uncertainty where the cost of being wrong is asymmetric.
That shows up in:
Climate risk, where tail events dominate outcomes.
Insurance systems, where rare events define solvency.
Financial markets, where regime shifts invalidate historical patterns.
Venture capital, where most signals are noisy but some are structurally informative.
I try to remain valid when stability breaks.
examples
Catastrophe and tail-risk modelling
Regime-switching dynamics
Scenario-based forecasting
Systemic risk analysis
Uncertainty quantification in complex models
I extract
structure from systems that don’t present it cleanly.
In practice, that means working with:
High-dimensional datasets
Time-series with structural breaks
Noisy or incomplete observations
Systems where the underlying process is partially hidden
Techniques I use include:
Statistical inference under uncertainty
Machine learning for pattern extraction
Latent structure identification
Simulation-based modelling
Monte Carlo methods
Data-driven dynamical system reconstruction
But the real focus is not the method - it’s the question:
What remains consistent when everything else is changing?
I build systems that connect theory to application.
That often means turning mathematical or statistical ideas into working tools - not just models on paper.
Examples of this kind of work include:
Simulation frameworks for dynamical systems
Operator-based computational pipelines
Automated decision-support tools
Data enrichment and intelligence systems (e.g. VC deal sourcing pipelines)
Interactive visualisations for complex behaviour
Forecasting and classification systems
Technically, this spans:
Python, Julia, R, SQL, MATLAB, JavaScript, AWS, and data visualisation tooling.
But the unifying theme is simple:
turning complex systems into something you can interrogate, test, and reason about.
The value is rarely in having more data. It’s in knowing what matters.
A useful test of any model is whether it survives contact with real decision-making.
I’ve worked across venture capital, actuarial science, and large-scale operational systems, where abstract models meet constraints like time, capital, and human judgement.
That includes:
Evaluating early-stage companies under uncertainty
Building deal-sourcing and intelligence systems
Structuring risk and return frameworks for investment decisions
Translating data into actionable investment insight
Improving operational decision pipelines in high-volume environments
Most systems look stable right up until the moment they aren’t.
Across mathematics, finance, climate, and venture capital, the same idea keeps appearing:
My work is ultimately about understanding that transition point - where stability ends and something new begins.
That’s where the mathematics becomes useful.
And that’s where decisions actually matter.