Canonical page: https://kevinmneal.com/

# Kevin M. Neal, PhD

Applied AI engineer & evolutionary biologist

Principal Engineer at Fortegra, with a PhD from UCLA. I build dependable AI systems and pursue independent research in biogeography and computational mathematics.

## Work & projects

### Fortegra

Principal Engineer

At Fortegra, I build applied AI systems for insurance workflows, including document-heavy underwriting.

My work spans architecture, document understanding, LLM workflows, evaluation, and production engineering. A recurring challenge is turning complex source material into structured information that people can trace back to evidence and review.

I previously worked on applied AI and document processing at Precedent.

### Spadefoot biogeography

Climate, distributions & conservation

My doctoral research explored how genetic variation and environmental differences inform spadefoot conservation. I’m returning to those questions through models of where *Spea* are recorded. The public project compares climate representations and their geographical predictions, with a working manuscript, released results, and reproducible comparison tables.

[Research & results](https://github.com/kevinmneal/spea-climate-research) · [Manuscript](https://github.com/kevinmneal/spea-climate-research/blob/main/docs/manuscript.md)

![Spea intermontana modeled recording density.](https://kevinmneal.com/assets/spea-intermontana-map.svg?v=20261015-rocket)

### Six triangles

An optimal packing, with replayable certificates

An independent computer-assisted optimality proof for a previously known arrangement of six unit equilateral triangles in an equilateral container. The project pairs exact, replayable certificates with an interactive explainer. Lean formalization is in progress; the full proof is not yet formalized in Lean.

[Interactive explainer](https://kevinmneal.github.io/six-triangle-packing/) · [Code & proof](https://github.com/kevinmneal/six-triangle-packing)

Six unit triangles

Other experiments: [LLM watermarking explainer](https://kevinmneal.github.io/llm-watermarking/) · [Lab & games](https://kevinmneal.com/lab/)

## A bit about me

I trained as an evolutionary biologist and moved into data science and then applied AI engineering after my PhD. I enjoy the work, but I never lost my interest in open-ended research. Building a career and raising a family left little time to pursue it on my own.

Recent reasoning models, paired with agents that can run code, inspect results, and keep working through a problem, have changed what I can take on.

What first caught my attention was the ambitious games people were building with AI. I made one too: *MAKE WAY!*, a medieval physics playground. Then I started wondering: could these tools do fresh work with my old PhD data?

Seeing AI-assisted mathematical research and formal proofs encouraged me to push further. That curiosity has brought me back to spadefoot biogeography and, unexpectedly, into computational geometry.

That’s what draws me to agentic science: opening up research I couldn’t otherwise find time for, while keeping evidence, uncertainty, reproducibility, and independent criticism at the center of the work.

## Publications & writing

[Scholar](https://scholar.google.com/citations?user=L6ln8EwAAAAJ&hl=en)

### Peer-reviewed articles

- [Conservation Genomics of the Threatened Western Spadefoot, *Spea hammondii*, in Urbanized Southern California](https://doi.org/10.1093/jhered/esaa049) — Neal, Fisher, Mitrovich & Shaffer · *Journal of Heredity* · 2020

- [Genetic structure and environmental niche modeling confirm two evolutionary and conservation units within the western spadefoot (*Spea hammondii*)](https://doi.org/10.1007/s10592-018-1066-7) — Neal, Johnson & Shaffer · *Conservation Genetics* · 2018

### Manuscripts & preprints

- [Climate representation and geographical prediction of recorded *Spea* occurrences](https://github.com/kevinmneal/spea-climate-research/blob/main/docs/manuscript.md) — Working manuscript · 2026

- [A computer-assisted proof of the optimal packing of six equilateral triangles](https://github.com/kevinmneal/six-triangle-packing/releases/download/v1.0.0/six-triangle-packing-preprint-v1.0.0.pdf) — Computational preprint · 2026

## Professional and scholarly profiles

- [GitHub](https://github.com/kevinmneal)
- [ORCID](https://orcid.org/0000-0003-4360-8027)
- [ResearchGate](https://www.researchgate.net/profile/Kevin-Neal-5)
- [LinkedIn](https://www.linkedin.com/in/kevinmneal/)