Tom Sullivan

AI & Data Advisory

I build data and AI platforms that have to be right.

Fifteen years turning enterprise data and AI from strategy into shipped, governed systems - across financial services, insurance, and life sciences. I also write about the part nobody warns you about: where agentic AI quietly breaks in production, and how to engineer around it.

Drawn by code, not generated by a model: every figure on this site is a deterministic render - the same discipline the essays argue for, applied to the essays themselves.

Flagship essay

Forward-Deployed Engineering Is New. The Work Underneath It Is Not.

Before forward-deployed engineering had its current label, implementation engineers and consultants were already walking operations, translating domain rules, and making software survive contact with reality.

2014 The year this work meant walking a regulated manufacturing floor and translating operations into validated software - before the FDE label.

Read the essay · the Field Guide maps the series · the reading list is the literature it stands on

Open source

Tools that make the argument executable.

Open source · npm · MIT

model-eol

A machine-readable model deprecation format and zero-dependency CI checker. It finds model IDs in a repository and makes retirement dates visible before they become emergency migrations.

Try it

npx model-eol . --days 90

Track record

$317MM+ value delivered - cost reduction, automation & ML
75+ engineers led across enterprise teams
1.1M documents a year processed at 80% lower cost
Tom Sullivan

About

I'm Tom. I've spent fifteen years building data and AI systems in industries where a wrong answer costs real money - pharma, consumer finance, insurance, and now private equity. The question that runs through all of it: not "can AI do this?" but "can you trust what comes out enough to act on it?" This site is where I work that question in public. The writing and views here are my own.

  1. Now

    AI & data advisory - private equity

    What AI and data are actually worth before a deal closes, and production systems that hold up after.

  2. Before

    Data & AI platform lead - Group 1001 (insurance)

    Led the platform teams behind the group's document AI and data estate - ~960k orchestrated production jobs a year at a 99.4% success rate, 1.1M documents a year processed at 80% lower cost. The transfer-automation work won the 2025 Gartner Eye on Innovation Award.

  3. Earlier

    ML platform lead - Discover Financial Services

    Led the build of a high-performance ML platform - three teams, ~40 engineers - that accelerated model deployment by 60% and supported 50+ models projected to drive $250MM+ in three-year profit before tax.

  4. Earlier still

    Automation organization lead

    $52MM in annual cost avoidance across an enterprise automation portfolio.

  5. Foundations

    Georgia Tech & Loyola Chicago

    MS in Computational Analytics; BS in Biophysics (cum laude).

The fastest way to reach me: tom@tomsullivan.dev or LinkedIn.