Work

Three platforms I built or led inside AT&T. Each one started with noticing something the org wasn't solving, and ended with other teams running their work through the thing I built.

Quickbase OpsHub

Built the platform an org didn't know it needed.

The situation

I was tasked with building a QA team for an Agile Release Train at AT&T. No one in the org had actually built a QA team here before. As I started, I noticed something bigger than the QA problem. People across multiple teams were running their work through spreadsheets and Word documents. Multiple copies of the same file. No shared system of record. Defects and enhancements were tracked in inboxes.

What I did

I asked permission to build a Quickbase instance. I picked Quickbase deliberately because the org already had a few apps in the Quickbase ecosystem, which meant my pitch would clear procurement easily. Then I went and got certified in Quickbase Foundations so I could actually build it. I shipped an app called OpsHub. It started as the QA team's tool. It quickly consolidated several other teams' spreadsheets. And then it became the intake system for the entire org's defects and enhancements, with KPI dashboards, UAT tracking, and a public defect reporting portal.

What changed

The org went from siloed file based tracking to a shared system of record. I went from "build a QA team" to "own the platform multiple teams now run through." None of it required a single direct report. The leverage came from the platform, not from headcount. The platform now anchors operations across PI cycles.

Ember Stitched Solution

Solved the front end nobody had built.

The situation

A few years into my AT&T tenure, I was the Ember.js front end expert on a 12 app enterprise portfolio. Two product owners came to me with components they wanted built for our customer facing apps. The components had never been done in the stack. They knew it would be hard. I told them upfront I'd have to come up with a unique solution. They smiled.

What I did

I went looking for the pieces. Three different plugins plus custom JavaScript could plausibly combine into the shape they needed. None of them on their own would do it. I tinkered in real time, made a lot of errors in the first hour, kept stitching. The first working version was ugly. Then I went back and cleaned it up.

What changed

The components shipped. The pattern of "stitch, ship, then refactor" became repeatable for similar requests. The product owners stopped asking other engineers whether something was possible and started asking me how I'd do it. It also became one of 100+ custom front end components I built across that 12 app portfolio (which served 5,000+ users at AT&T).

Agentic AI Test Case Generation

Production AI inside QA scope.

The situation

QA test case authoring is one of the slowest and most repetitive parts of the engineering pipeline at scale. Existing tooling either generates cases too generically to be useful or requires so much human authoring that it doesn't change the velocity equation. I'm currently QA Lead for one Agile Release Train at AT&T (this is current scope, not org wide).

What I did

I built an agentic AI workflow that generates test cases from requirements inside the QA scope on my train. Production tooling, not a slide deck. The workflow is the thing that runs, not a future "we should explore AI" plan.

What changed

Eliminated 17 to 19 hours of manual test case authoring per PI, with 85% of generated cases usable without edits. Built on Microsoft Copilot plus Quickbase. More importantly: I have a real artifact of building inside the new AI paradigm rather than just talking about it. That changes the conversation in interviews from "what do you think about AI" to "let me show you what I built."

Selected work.

Eastern Poultry Distributors ERP (2013). Built a complete ERP, warehouse management, and CRM platform from scratch in PHP, MySQL, and CodeIgniter for a regional distributor.

South Georgia Medical Center first iPhone app (2014). Delivered the organization's first iPhone application (Xcode/Swift), a new patient facing channel for a regional medical center.

Look Into The Well, LLC (ongoing partnership). Sole developer for a small Georgia branding agency. I built and maintain the agency's own site (lookintothewell.com), and I build client websites for them from designs handed off by their design team. Selected client projects:

  • UPPERCUT Homes (uppercuthomes.com). Built site for an award winning home builder.
  • Griffin Historical Society (griffinhistory.com). Built site for a historical society redefining mission and brand.
  • Lighting Associates (lightingassociates.com). Assisted on site through a successful merger of an industrial lighting agency and a design firm.
  • OutdoorLink. Assisted on site for a wireless outdoor lighting control company expanding into new verticals and reseller programs.

Custom Livestream WordPress Plugin (2026). Production WordPress plugin built with Claude Code that gives church admin staff control over livestream scheduling, on page content, and countdown display from a single widget.

Client Monitoring Portal (2026). Centralized self use web app built with Claude Code that monitors all public facing client sites, surfaces status and downtime alerts, and supports remote VM reboot.

Consulting and client work.

A short list of clients through WTX Labs. Full portfolio at wtxlabs.com.

  • Eagles Landing
  • Colson Dental
  • Stewardship Journal
  • Oasis Counseling
  • Bridgepoint Toco