
Above the Cloudline
The clouds hold the philosophy: plan practically, dream dramatically — close enough to the vision to follow it, grounded enough to build it.
Agentic software engineer
Previously front-end tech lead at The Home Depot; now co-founder & CTO of SmrtFood — instrumenting AI agents and shipping production React at enterprise scale.
I build the interfaces the rest of the company quietly relies on — the dashboards that route a delivery fleet, the scanner that double-checks a child's lunchbox, the admin tool that lets ops mark a warehouse closed before the next truck rolls. The visible surface is mine; so is the contract underneath it.
Six years of production React in enterprise contexts — most of it as the front-end lead for The Home Depot's Delivery Network Admin tool, the rest co-founding and shipping a React Native consumer app that started as a question over my kitchen table. I'm most useful where craft and ownership matter: where a team needs someone who can write the spec, draw the component boundary, push the migration through, and still care about the kerning when it lands.
Lately that craft extends to the agents working alongside me — Claude Code, Codex, Antigravity MCP — which I instrument like anything else in production: a hook ships each subagent's transcript to Gemini Flash, which reports back to my private Discord as the agent itself, one persona apiece. The fleet checks in; I read it over coffee.
Rather than stock backdrops, these are visual studies I generated for this space — the same agentic tools I run at work, turned on the craft itself. They hold one balance: vision expansive enough to inspire, execution sharp enough to build.

The clouds hold the philosophy: plan practically, dream dramatically — close enough to the vision to follow it, grounded enough to build it.

Each star is a facet of the vision; each line turns it into a plan. The clouds clear inside the mirror as the dream becomes something solid enough to act on.
Led front-end development of the Delivery Network Admin tool — an enterprise React application used by 2,400+ Home Depot distribution centers to manage warehouse capacity, shipping routes, and delivery scheduling. Acted as Epic Champion: technical specs, database design, cross-team coordination, sprint planning.
End-to-end ownership: React UI, Spring Boot APIs, Pub/Sub contracts, PostgreSQL schema.
Adopted by 2,400+ buildings in production, eliminating shipping delays caused by manual capacity calls.
Migrated an Angular PoC to React / TypeScript / Next.js / Node.js across 20+ feature modules.
Performance and maintainability lifted across the board. Defined the security framework (PingFederation SSO) on the rewrite.
PostgreSQL ↔ BigQuery sync giving DNA users self-serve market lookups.
Eliminated repetitive engineering inquiries by surfacing the data behind the questions users were already asking.
Dynamic delivery-slot allocation with predictive demand modeling.
Architected the allocator and integrated the prediction layer; visibly improved scheduling efficiency and on-time rates.
Built the initial DNA tool as a Proof of Concept (Angular + TypeScript + Ant Design). The PoC secured full project funding and the team expansion that became the next chapter.
Took the original Angular proof of concept from sketch to a working tool stakeholders approved.
The PoC's reception is what built the team I'd later lead.
Domain services in Java Spring Boot + PostgreSQL; full JUnit coverage.
Set the backend conventions other DNA engineers built against.
Established unit, integration, and end-to-end coverage with Jasmine and Cypress.
Cross-functional work with Product, Infrastructure, and Architecture teams.
Four products I'm actively building outside the day job. Each set in its own brand voice — colors, type, and tone run independent of the surrounding magazine.
Career intelligence that lives on your machine — not someone else's funnel.
A local-first, autonomous job-application companion for senior operators. Reads the role, the rumor, and your work history; drafts the application, queues the follow-ups, and keeps every byte on-device. The opposite of mass-apply.
Teach your dog in any language — and meet them halfway.
A foreign-language dog-training companion. Pick a target language, learn the cue, practice the pronunciation, and run training drills your dog can actually follow. Built for owners who want their working breeds to listen, regardless of which side of the border they grew up on.
A scanner that double-checks a child's lunchbox before someone else has to.
A React Native barcode scanner for parents with children in daycare or school food programs. Scan a food item — the app pulls nutrition info from the FDA database, cross-references the ingredients against every child's dietary profile, and flags allergens or restrictions before the food gets handed off.
The AI co-pilot — "Crewmates" — that makes a one-person stream feel like a room.
A streaming co-pilot for new creators. Crewmates ride along with the broadcast, reading chat, riffing on the moment, and filling the dead air a brand-new streamer can't yet fill on their own. Landing page is live; the dashboard scaffolding is in.
Six disciplines the work has actually run on. Production-grade, not buzzword-grade.
Open to senior front-end and tech-lead roles where craft and ownership matter. Available in Atlanta or remote-first, starting in 2026. Reply within 24h.