The Whole-Stack Shift
AI didn't replace the engineer. It widened the engineer.
For most of software's history, the whole production stack was too much for one person. That changed. One engineer can now hold all of it, across software, data, AI, cloud, security, and delivery. Whether you are building a new product, delivering a program, or facing a delivery in trouble, I go in hands-on, lead it to production, and leave a team that can run it.
Behind every software delivery and rescue I take on, there is a system
Bringing AI and ML into real software systems, new and legacy alike, and making them hold in production. That is the work. What makes it possible is the system behind the work: fifty years of how the industry actually builds production software, plus thirty of my own in the trenches, packaged into one method. Software, data and AI engineering, DevOps, SRE and observability, and the security and compliance that a chatbot or a shallow tool quietly gets wrong, and often makes worse. It grows sharper with every production system I deliver. Sign up to follow it, or book a call to see it live.
A new piece as it lands. Unsubscribe anytime.
The shift
Point AI at the work and it produces plausible versions of each piece: a data pipeline, an ML service, a multi-agent RAG system, an auth layer, a cloud setup. But plausible isn't a working product, and none of it holds together on its own.
What makes it a real system is engineering: real depth across the disciplines, a proven end-to-end template that covers them, and the judgment to make the pieces hold. That used to be an org chart. Now it's a skill set. The failures live in the seams between the disciplines. A relay of specialists hands those seams back and forth. One whole-stack engineer owns them.
It takes method, not heroics. A controlled, spec-driven build with defined agent roles, one agent checking another, and the discipline to catch the slop. Underneath sits one skill: asking the right question at the right depth, probing until the real gap shows. Same muscle whether you are reading a system, a team, or a prompt. It is how I turn projects around, and how I get the most from an agent.
It's wide and deep at once. Width used to cost depth. You picked one. AI removes that choice.
The wall came down. A great idea used to need funding, a team, an org chart. Now one engineer, or a small team, can build what used to take a department. The constraint was never talent. It was cost.
The widening is not only technical. It is becoming someone who can decide what is worth building, then design it, build it, and ship it. When ideas are cheap to try, knowing which one to build is the real edge.
AI holds the promise, but it doesn't deliver it. Closing the gap between what it seems able to do and what actually ships is the whole game. It's a muscle you build, not a model you buy.
Why this holds
Ask any AI for a checklist or a plan and it will give you one. What no AI can give you is the system that makes the work hold in production.
Half a century of how production software is actually built, packaged into one method. Software, data and AI engineering, DevOps, SRE and observability, and the security and compliance that superficial tools get wrong. Sharpened on every system I take through to production. Depth across every layer, not a thin tool over one. That is what makes the result still stand after I am gone.
Work with me
The AI build that impressed the board, then broke on real data. The delivery a week from its deadline with no one who can own the whole system. That is the work I do. A commodity shop hands you a build and moves on. I go in hands-on, lead the engagement end to end, and leave a proprietary way of working that keeps delivering after I am gone. Delivered company to company through Emergtech.
Rescue a delivery in trouble
I turn dysfunction into delivery momentum.
An AI system that passed its demo and is failing in production. A client on the verge of leaving. A critical delivery on fire. A launch about to slip. I come in hands-on and get it moving again. But the fire going out is the smallest part of it. You come out with a team that delivers reliably on its own and works confidently across the crossing, the AI and the production engineering both, with the methods and judgment to keep it that way, yours on every project after this one. The techniques I spent 30 years building in the trenches become yours. And I do not disappear when the fire is out. I stay for what comes next.
Build it end to end
A product you own, and the way to keep building.
Maybe you already have a clear product idea, or just a domain you know well and a sense there is something worth building. We start there and I build it with you, end to end and across the whole stack. You come away with more than a product. You get the templates, the method, and a team that can run it and reason confidently across the whole stack, AI and traditional engineering both, so the next build is faster and the one after that faster still. Not a one-time delivery, but a capability you own for good, with a partner who stays as you grow. You get the senior engineer you could not hire full-time, for exactly the build in front of you.
Selected work
A few engagements, anonymized. The pattern repeats: go in, make the pieces hold, leave a team that can run it.
A critical delivery, a week from the penalty clause
A mission-critical reporting platform was a week from the client invoking its penalty clause, with the reports the business ran its decisions on coming back wrong and a team with no room left to think. I bridged the teams to one ground truth, then redesigned the system while it stayed in production, using every ticket to fix the foundation beneath it. The flood slowed to the ordinary load a healthy system carries, and the client stayed.
A turnaround is two jobs: stop the bleeding, and leave behind something that does not bleed again.
A $100M program, and “it is just a simple system”
A $100M platform nearly failed because leadership saw only the surface and chose an off-the-shelf replacement, not the years of business logic, the dozens of data sources, and the edge cases that had quietly become load-bearing walls. The replacement became a black box that fought every customization. I have been part of the teams brought in to rescue programs like this, finding every crack where you can inject a bit of sanity.
The complexity is never in the tool. It is in what you are replacing. The leaders who get it right ask early: what are we not seeing?
A million lines, migrated in one coordinated pass
A mission-critical platform had been stuck for three years. About a million lines of code, five languages, several databases, and dozens of integrations, all too entangled to take a piece at a time, with the clock running out. Programs like this stall when no one holds the whole. It moved when the team stopped guarding boxes and asked what the work actually needed.
Done in eight months, the kind of migration that usually takes years. One attempt. No fallout.
What I bring beyond the build
Beyond the build, I bring a system of my own. Proprietary methods and accelerators I have developed for production reliability, deployed inside every engagement. It is the difference between a demo that impresses and a system that holds when the real data hits. A checklist tells you what to check. This is what makes the check pass. And because the method stays with your team, the next build is faster and the one after that faster still.
What people say
From the leaders, architects, and managers I have delivered alongside.
“A committed individual who does not rest until he gets the problem solved, be it an outage or a Sev 1 type of situation. He constantly gets accolades from the clients and is there with the team when they need him the most.”
“A technical expert who is assigned to the toughest problems the team faces. An expert technical leader and troubleshooter who embraces new technology and can develop across numerous technologies with great success.”
“More of a thinker and a doer and less of a talker. It was his due diligence as an architect that made the first large application a resonating success.”
“Always instrumental in clearing critical issues and delivering successful releases for a complex, business-critical platform. His ingrained ability to find the root cause and fix the problem helped us immensely.”
“A great leadership asset for any organization, whether hyper-scale growth or troubleshooting complex situations. He builds high-performance teams in the shortest time and gives them the momentum to hit time-bound objectives.”
I build new production systems and rescue ones in trouble, from a small team's first launch to a $100M program a week from collapse. AI raised the ceiling on both. Rescues matter more now, as teams bolt AI onto systems no one fully understands. And building a whole product end to end, across software, data, AI, cloud, and delivery, is faster and more exciting than ever, within reach of one person or a small team.
Library
In-depth work on building and rescuing production systems in the AI era.
The ebook
Why one person or a small team can hold the whole stack
The full series compiled into one read: the method that makes it possible, then the layer tour from getting data in to shipping it to production. Get it the day it is ready.
A new piece as it lands. Unsubscribe anytime.
The checklist
Will it survive production?
A readiness checklist for AI systems. The things I check when I go into a build or a rescue.
Read the checklist →Ready to hold the whole stack?
Book a call and we'll figure out the right way to build together, or start with the ebook and the field notes.