AI Research Scientist | Systems Designer
Understanding the foundations. Discovering the connections.
My independent research explores how understanding the underlying structure of complex systems can reveal new solutions, connect discoveries across disciplines and improve future problem-solving.
Conviction. First Principles. Discovery.
I believe the most important advances begin by questioning what we think is possible.
My work is driven by a conviction that the foundations of complex problems can reveal opportunities conventional approaches overlook - and that discoveries should improve not only individual solutions, but the process of discovery itself.
I develop and test systems designed to make that possible.
The Foundation
Understanding systems from their deepest foundations.
Originally developed to understand how powerful stories and coherent worlds are built, The Foundation evolved into a common, fractal structural language for describing systems.
It is designed to represent components, relationships and interacting systems at different scales, making seemingly unrelated disciplines comparable without losing what makes each unique.
This provides a foundation for understanding how systems work, diagnosing limitations, improving existing designs and creating entirely new ones.
Structural Intelligence
Discovering the principles that connect systems.
Structural Intelligence builds on The Foundation to identify consequential patterns, invariants, hidden constraints and governing principles across different systems.
Its purpose is not simply to find similarities, but to discover when knowledge from one problem can genuinely help solve another. Proposed connections must be tested against the mechanisms and limitations of each domain.
By retaining verified discoveries, failed approaches and their boundaries, the aim is to progressively improve how unfamiliar problems are understood and approached.
Research Evidence
I deliberately selected unfamiliar mathematical and computational domains with established baselines and objective correctness checks, allowing the research to be judged by its results.
Qualified Structural Memory
20/20 reference routes · 5.31× lower selection burden
Cross-domain holdout results comparing retained structural knowledge against cold routing and surface-result memory.
Cross-Domain Structural Transfer
4 computational families
A shared reduction strategy produced selected, correctness-checked results across Markov, stochastic, automata and linear-programming problems.
Numerical Optimization
1.59× Netlib · 2.10× Kennington
Enki + HiGHS performance in recorded internal comparisons against default HiGHS across 108 linear-programming benchmarks.
Specialist Routing
12 specialist designs
Research into selecting effective problem-solving methods by structural characteristics, with implemented components evaluated against generic baselines.
Independent research. Experimental methods, conditions and limitations are provided in the detailed evidence pages.
Collective Excellence