Senior Data Engineering · AI-Assisted

Your systems already know. I make them tell you.

I'm Will Jewell, in Chattanooga. Your business runs on software — accounting, scheduling, dispatch, whatever you use day to day. I get the information out of those systems, check it against what it should say, and turn it into something you can act on: a number you can trust, a small tool your team actually uses, or a flag when something needs attention before it becomes a problem.

Email me the problem See how I work

Getting data out · of systems that don't share it Joining it up · across tools that disagree Checking the numbers · against what they should be Catching what's wrong · before anyone else notices Reports that hold up · when someone checks them Keeping it running · without me babysitting it

12+ yrsdata engineering

180+hospitals served

~20Kdaily users fed

76→97%data quality lifted

Usually the answer already exists. It’s just stuck.

Your scheduling software knows half of it. Your accounting system knows the other half. Neither was built to talk to the other, so the questions that would actually tell you something — which jobs lose money, who pays late, what’s about to go wrong — are the ones nobody can answer without a week of spreadsheet work. So they don’t get asked.

I get the information out of each system, line it up so the pieces match, and fill in what’s missing around it. After that, those questions become ordinary ones you can ask any time you want — and you keep everything I build.

What I do

Most of my work is one of these six things. The messier your information is, the more useful I am.

The messier the data, the more useful I am.

Getting your systems to agree

Your numbers live in three or four places that don’t match. I pull from all of them and produce one version everyone can work from.

Knowing now, not tomorrow

When something needs a same-day response, the information has to arrive the same day. I set that up.

One place the answers live

Instead of five spreadsheets and someone’s memory, one place to ask a question and get the same answer twice.

Reports that hold up

Right underneath, not just tidy on top. If a number looks wrong, you can trace where it came from.

Fast, and checked

I use AI heavily to build quicker — then test it and show the working. The checking is the part AI doesn’t do for you, and it’s the part that matters.

Getting data out of software that won’t share it

Plenty of tools hold your information hostage with no clean way to export it. I get it out anyway, and keep it flowing without anyone re-typing it.

How I work

Production is the finish line — not a slide deck.

A tight, transparent loop. You see working software early and own it at the end.

Frame

Understand the data, systems, and decisions it feeds. Scope the smallest useful thing.

Build

AI-assisted where useful, but still versioned, tested, and inspectable — not a black box.

Ship

To production, monitored, with alerting when something drifts.

Hand off

Documented, so your team owns it. No black boxes, no lock-in.

About

Twelve years building the plumbing behind other people's software — the part that moves information between systems and checks it's right. Mostly in places where being wrong has real consequences.

Right now I run data for clinical trials at Thermo Fisher, making sure teams building software and AI tools are all working from the same, correct set of numbers. I also work out how many patients would actually qualify for a study before the company bids on running it.

Before that I spent years at HCA Healthcare keeping the data flowing behind an app that about 20,000 clinicians used every day across more than 180 hospitals. If it broke, doctors noticed within minutes — so I got very good at building things that don't break quietly. Earlier, I took a company's main database from 76% correct to 97%. Degree from Georgia Tech.

None of that is specific to hospitals. What keeps medical records straight is the same thing that keeps a contractor's bid package or a carrier's paperwork straight: knowing where every number came from and noticing when one is wrong. Whatever I build, you own it — written down, and running without me.

Outside work: being a dad, mostly — plus a growing pile of Raspberry Pis and a community fridge I help keep stocked. I especially like working with climate, health and civic teams, where getting the data right actually moves something.

If you’re in one of these industries

Three kinds of client kept asking for the same thing, so I packaged those at a fixed price. If none of them is you, ignore this — the work above is the same either way.

Two simple ways to work together.

Fractional retainer

An ongoing data-eng function

A set slice of my week, every week — for continuous data work.

Predictable monthly cost, no full-time overhead
Senior hands from day one
Scale up or down as needs change

Project / SOW

A defined build, shipped

A scoped outcome — migration, pipeline, or reporting layer — delivered to production.

Clear scope and deliverable up front
Shipped, monitored, documented
Your team owns it at handoff

Reference & write-ups

Things I’ve worked out, written down.

Mostly the specifics that are hard to find anywhere else.

A sample driver file review — the document a carrier gets back
A sample bid package review — what a federal solicitation turns into
A sample one-load review — the invoice package, worked through
Medical certificates after June 2025 — CDL drivers went electronic, non-CDL did not
The DOT driver file calendar — every recurring deadline, and what triggers it
Getting data out of AppFolio — the API is one-way and gated
Customer PO to Sage 100 — emailed orders into validated drafts

Start asynchronously

Got data that's more headache than asset?

Email me the systems involved, what keeps breaking, and what the output should look like. I'll reply with the first practical step and whether I'm a fit.

Email me the problem