Skip to content
Steven Purcell

About

Who is the Data Science Ninja?

My name is Steven Purcell. I'm a data and analytics leader with experience in enterprise data strategy, business intelligence, artificial intelligence, machine learning, and cloud-based analytics across healthcare, government, and commercial environments. I currently serve as Director, Data & Analytics at Catalight, and I'm the founder of Fruitful Intelligence LLC, my ongoing data science and technology consulting business. My work spans analytics team leadership, data platform modernization, predictive modeling, data governance, and technical solution delivery.

I have a Bachelor of Science from Austin Peay State University in Applied Mathematics with a minor in Physics, and a Master of Science in Data Analytics from Northern Illinois University. I am passionate about automation and cloud development.

When I am not performing impressive acts of Data Science Ninjutsu, I enjoy kayaking, hiking, backpacking, rock climbing, and 3d printing. I collect action figures and comic books. I also enjoy beer, the hazier, citrus-y, and doublier the IPA, the better. I live in Virginia Beach and I’m very keen on protecting the Chesapeake Bay. I spend my downtime reading sci-fi and fantasy novels, playing Pokemon, and collecting things. My GitHub is full of half-finished projects that I constantly promise myself I will get back to, but I am always distracted by the next recreational coding project. I consider myself a generalist and I enjoy learning new technologies.

Some projects I am proud of: I wrote an Android app in Dart and Flutter that tells dad jokes guaranteed to make you roll your eyes. I used Go to fully automate database migrations in AWS to save myself an hour each time we needed to update data in the cloud.

I do not like beets, I think they taste strongly of dirt and they make my teeth red. I once sarcastically pretended to check the time on a nonexistent wristwatch and inadvertently spilled coffee on myself. Once, for $5, I juiced a hot dog. I don’t know why I did it.

The path here

Not the most linear route into data science — which turns out to be useful.

  1. 2007 – 2012

    Emergency dispatcher → dispatch supervisor

    Cheatham County, TN. Where I learned to stay calm, document everything, and make decisions under pressure.

  2. 2016

    B.S. Applied Mathematics

    Austin Peay State University, with a minor in physics — and a year helping other students as a learning-assistance leader.

  3. 2016 – 2017

    Data analyst → systems analyst

    Credit and debit payments data at edo Interactive, then a healthcare data stream from roughly 170 hospitals at Community Health Systems.

  4. 2017 – 2024

    Lead data scientist → VP of Technology

    Built predictive models on AWS and led the analytics and engineering teams around them at G2Lytics, including the nForce sales-tax noncompliance detection application.

  5. 2021

    M.S. Data Analytics

    Northern Illinois University.

  6. 2023 – Present

    Founder, Fruitful Intelligence LLC

    My ongoing data science and technology consultancy: technical advisory and implementation support in AI/ML, analytics engineering, cloud solutions, and workflow automation.

  7. 2024 – 2026

    Adjunct faculty, SNHU

    Taught graduate-level data analytics courses online at Southern New Hampshire University (March 2024 – August 2026).

  8. 2024 – 2026

    Manager → senior manager, InDev

    Led AI/ML and advanced analytics work supporting U.S. Department of Homeland Security programs, including enterprise reporting modernization and cloud data integration (April 2024 – April 2026).

  9. 2026 – Present

    Director, Data & Analytics, Catalight

    Leading enterprise data and analytics: business intelligence, AI/ML, predictive forecasting, data governance, and analytics platform modernization for clinical, finance, product, and operations teams.

How I approach problems

Start simple, earn the complexity

A moving average that ships beats a neural network that doesn't. I reach for fancier methods when the data asks for them, not because they're shiny.

Traditional vs AI forecasting →

Context over cleverness

Most "AI problems" are really missing-context problems. Give a model the definitions, metadata, and constraints a new teammate would get, then check its work.

Context-aware SQL generation →

Keep humans where judgment lives

Automate the routine, surface uncertainty honestly, and route the genuinely ambiguous cases to a person with enough context to decide quickly.

ID validation →

Protect focus time

Fewer, better meetings. A small team with long uninterrupted stretches will out-deliver a busy calendar every time.

Lean management style →

What I'm into, technically

  • Data strategy & governance
  • Business intelligence
  • Machine learning & MLOps
  • Large language models
  • Predictive analytics
  • Fraud & anomaly detection
  • Forecasting
  • Python
  • Go
  • SQL
  • Snowflake
  • AWS
  • Docker
  • Linux (Mint, these days)
  • Automation

Looking to hire or partner?

This is my personal site. For consulting, contracting, or other business inquiries, please head toSoFruitful.com.