available · engineering leadership + applied ai · atlanta, ga
I change how engineering teams work.
Then I prove it.
I'm Antonio Yon. Across 25+ years I've led engineering and product engineering at Amazon, Cox Automotive, Sysco, and Phreesia, most recently at the director level. What I actually do is retire the processes teams have learned to work around: the bespoke integration built per customer, the escalation chain nobody trusts, the release cadence everyone plans around. Lately that work centers on AI in the software lifecycle. I built PRISM at Phreesia, and I now author SPECTRUM ↗, a governed agentic SDLC framework that lets AI agents write production code.
leading engineering
agentic SDLC I built at Phreesia
per-customer builds with reusable services
5 cross-functional teams
how i lead
Senior enough to set direction, close enough to the code to be useful.
I lead managers and teams, but I've never stopped being a builder. That combination is what lets me modernize platforms and bring new technology in — AI agents included — without stalling delivery.
Lead managers, not just engineers
Manager of managers. At Phreesia I owned five cross-functional teams, with direct reports that grew from eight to twelve once Product Engineering folded in. Earlier, I scaled an iOS org from a single office into multiple teams.
Bring AI into the SDLC, responsibly
I built PRISM at Phreesia and now author SPECTRUM ↗. The framework was never the hard part. Getting teams to adopt it was. I proved it on one team first and let the results make the case to the next, and it spread to eight teams, three of them outside my own organization. I advise teams on this separately through Antfarm Solutions ↗.
Modernize without stopping delivery
I move platforms to cloud-native footing while the business keeps shipping — Phreesia's integration stack onto AWS EKS, Sysco's internal platform onto OpenShift — sequenced by business risk, not technical preference.
Make delivery faster and safer
Operating-model and testing redesigns that cut Phreesia's integration cycles from two weeks to about two days, and a shared API and contract layer that other product teams built on instead of reinventing.
pace notes
What I've learned about leading
I'm a reflective leader by nature, and I've kept blogs built around exactly that — First by Reflection, and The Greater Fool, where a mid-career dive into economics had me writing about incentives and how not to be the greater fool in any deal. What's stuck with me most is the discipline of turning experience into principles. A few have held up across every team I've led.
"By three methods we may learn wisdom: first, by reflection, which is noblest; second, by imitation, which is easiest; and third by experience, which is the bitterest."
— Confucius
Psychological safety comes first
I build teams where every person is empowered to form and voice an opinion — empowered individuals, not just well-managed ones. A culture where every voice carries is the foundation everything else sits on.
Name the risk before you have the fix
Too many people only surface a risk once they've got a mitigation in hand. Saying "I don't know yet" out loud is part of leading, not a crack in it — accepting that we don't know everything is the job.
Take people where they wouldn't go alone
The leaders worth following move a team to places it would never reach on its own. That's the whole point of the role.
experience
Selected experience
full-time focus
Founder & Principal Advisor
My long-running advisory practice, now my full-time focus since Phreesia. I help software teams adopt agentic AI responsibly — AI strategy, governance, and agentic SDLC grounded in SPECTRUM. Recent work includes a build where two developers new to the codebase delivered nine stories in about a day and a half, against estimates a solution architect had written for senior engineers. I'm currently the full-stack engineer on a Large Action Model project that navigates and interacts with real third-party websites. antfarmsolutions.com ↗
Director, Network Solutions — Engineering & Product Engineering
Owned five cross-functional teams (~36 people; direct reports grew from 8 to 12 once Product Engineering folded in). Built PRISM, an in-repo agentic SDLC framework that captured specs, BDD tests, and architectural decisions so AI coding agents had durable context and every decision had a reviewable history — adopted by eight teams. PRISM was internal to Phreesia and stayed there. SPECTRUM is the framework I built independently afterward and put in the open. Also redesigned the operating model to shared ownership across product, architecture, and engineering.
Director, Integration Engineering
Led five teams running Phreesia's national integration platform (API, HL7, FHIR). Retired per-customer bespoke integration builds in favor of reusable domain services with thin adapters at the edge, which consolidated 12 integration teams into three organized by integration type. Migrated the stack to cloud-native AWS (EKS, SQS/SNS, DocumentDB, DynamoDB) and set the API and contract standards other teams built on.
Senior Architect
Defined target-state architecture and migration paths — API-first contracts, event-driven flows, observability, clear security boundaries — balancing risk and partner impact over an all-at-once cutover.
Engineering Platform Owner
Built the internal developer-productivity platform on Red Hat OpenShift, then golden paths that let product teams adopt managed Kubernetes without learning Kubernetes first. When COVID closed restaurants, my team shipped a direct-to-consumer marketplace end-to-end in two weeks — the platform is what made that velocity possible.
Principal Solution Architect
Architected and hands-on built a customer-facing deal widget embedded on Dealertrack and Autotrader vehicle pages (React, Node.js, GraphQL, AWS) handling credit check, trade-in, and price negotiation. Promoted from Solution Architect to Principal.
Senior Technical Program Manager
Drove experimentation strategy for the amazon.com A/B testing platform, raising the bar on experimental design and on what counted as a real result rather than a launch ceremony.
Earlier engineering & architecture leadership
Architecture and engineering leadership across fintech, healthcare, and consumer services — API aggregation, commercial lending on .NET/Azure, native iOS teams, and utility billing platforms.
where engineering is going
I'm already working where most orgs are headed.
Most teams are still figuring out how to let AI agents write production code responsibly. I've spent years on exactly that problem. First with PRISM inside Phreesia, which stayed there when I left, and then with SPECTRUM ↗, which I rebuilt in the open.
SPECTRUM ↗ is my governed, spec-driven framework: specifications, BDD tests, and architectural decisions captured in-repo so agents have durable context and every decision has a reviewable history. It's the kind of forward investment I bring into the organizations I lead.
I also built ragpak, a service-metadata and retrieval layer that lets engineers and agents reason across an organization's services when debugging across boundaries. ASP.NET Core over gRPC to a Python sidecar, PostgreSQL with pgvector and a graph store, on Kubernetes. It was built using SPECTRUM, which is the proof the framework holds up on something other than itself. Ingestion, provenance, and the HTTP contract are tested; the retrieval query path is the next iteration.
work with me
Bring me in to make AI a first-class part of how your teams ship.
Through my advisory practice, Antfarm Solutions ↗, I help engineering organizations adopt agentic AI without giving up control of quality, security, or accountability.
Stand up a governed agentic SDLC
Adopt SPECTRUM ↗ in your repos, with the governance model and human review checkpoints tuned to how your teams actually ship.
AI strategy & governance
Where AI actually pays off in your delivery pipeline, and the guardrails (security, accountability, human review) that make it safe to move fast.
Hands-on builds
I still build. Recent work includes a Large Action Model that navigates real third-party websites — full-stack, shipped, in production-grade shape.
Get teams to actually use it
The framework is the easy part. Most AI initiatives die as a pilot that never spread. I prove it on one team, make the results visible, and let the next team come asking.
let's talk
Let's build something.
I'm open to senior engineering leadership roles — Director, VP, or Head of Engineering — based in or around Atlanta, and I take on a small number of advisory engagements through Antfarm Solutions. If either is what you're looking for, I'd like to hear from you.
or write directly to antonio@antonioyon.com