Harlan T Wood · AI & software
25 years building software. Now I help mission-driven people and organizations build with AI — powerfully expressing their vision with integrity and heart.
“If Harlan Wood is seeking to work his magic for you, and you decide to go forward, simply put, you will never regret it… vast technical talent, extensive creative gifts, and the discipline and impact of a seasoned master.”
You can feel the pressure to do something with AI. You can also feel when the options in front of you have no soul.
You can feel it pull in both directions — and you don’t have to choose. This is not hesitation. It is discernment. It also happens to be good engineering instinct: the AI systems built without that question are, reliably, the ones that fall apart six months into production.
You can move quickly and build something you’d be proud to have made. I’ve been there before, and I’ve spent twenty-five years learning how to do both at once.
Three ways I help, whether you’re starting from a blank slate or a system that has begun to drift.
Your idea or prototype, built to become the actual product — security-hardened, productionized, and architected to scale. Not a demo you have to throw away when it is time to get serious, but the foundation you keep building on.
I review your existing software and bring it up to production grade: security, hardening, scalability, and the tests and automation that keep it healthy. You are left with systems your own people — or an ordinary developer with AI assistance — can confidently maintain and extend.
Set the whole thing up from scratch — the team, the tools, and a task-driven pipeline you can largely forget about, where AI does the heavy lifting and people steer. I love doing this with startups: it is the difference between shipping massive value weekly that used to take a quarter.
You don’t need a polished AI brief before we talk. Bring the ambition, the napkin drawing, the half-working prototype, or the system that has begun to drift. We start from what’s true.
We map the decisions, constraints, and human consequences. You leave with a grounded technical direction and a clearer sense of what deserves to be built.
I design the architecture and work directly in the code: agents, voice, evaluation, retrieval, and the operational infrastructure that makes it all dependable.
I pair, teach, and upgrade the working habits — and the systems around the software product itself. The aim is a team that is fully AI-empowered after my part is done.
“Harlan is great at mobilizing and managing teams, as well as interacting with the non-engineering folks in a company, getting both groups into alignment and onto the same page.”
Not as a hobby, and not in hindsight. Each of these was an unfashionable place to spend a year of your life at the time.
Twenty-five years of hype cycles teaches you to tell the difference between what is arriving and what is merely loud — and to know that the two are rarely the same thing.
What’s arriving now is real. Most of what’s loud isn’t.
“Harlan was a progressive force in the company, promoting agile practices and helping us improve the way we made software.… More than anything, Harlan brings a passion for developing software the right way.”
Recent client work.
Zero to production: products taken from first commit to security-hardened systems with real users depending on them — architecture, build, deployment, and the operational discipline that keeps them standing.
I design evaluation frameworks and introduce eval-driven development to teams as a core practice — custom simulators running matrix comparisons across models and runs, wired into the tooling developers already use. This is the single highest-leverage thing most AI teams are missing, and the reason so many promising projects stall.
Voice-to-voice systems built on VAPI, which lets us use any combination of models for transcription, voice synthesis, and the LLM brain — tool calling and custom webhook handling, tuned until the latency stops being something you notice. Getting a model to answer is easy. Getting it to feel like someone listening is the work.
A knowledge-graph system that generates nodes and edges from natural-language queries, powering a path-based retrieval architecture that outperforms both conventional RAG and GraphRAG. Twenty-five years of studying semantic graphs, arriving exactly where the field now needs it.
HIPAA-compliant applications, WASM for offline capability, content security policy, CI/CD that developers don’t fight — and that raises quality through AI automation. An AI product is still a software product, and building it like one is what lets it thrive.
“Harlan is one of those too-rare people who can see what needs to be done, how to cut the problem down into bite-sized pieces, and then knock the code out of the ballpark. Both a pragmatic programmer and an agile project manager, I’d work with him again on any project that needs a guy who makes things happen!”
These are the long projects — the ones I’ve given years to because I want them to exist. They are also the most honest description of what I think technology is for.
A 3D social operating system and regenerative impact dashboard — a memory palace for your own self-sovereign data, rendered as node sculptures built for you by AI while you sleep, in the spare cycles of your idle devices. And that’s only the start: a shared world for presence, play, and collective sensemaking.
Depth-psychological, archetypal astrology in a luminous, hand-crafted interface — AI doing the synthesis while the voice stays human. Its palette quietly inspired this very page.
An open-source interchange format for reputation. Neutral and unopinionated: a base layer that lets decentralized systems share trust selectively. Built so that peer-to-peer credit ratings for the unbanked, refugee credentials, and decentralized peer review become actionable and simple, rather than remaining good intentions.
Fractal royalties for all beings. A proof-of-concept built on Holochain.
The monolithic towers of the tech monopolies are giving way to something more like mycelium — lateral, redundant, cooperative, rooted. Not because the towers lost an argument, but because the other shape turns out to carry more life.
What I care about is the difference between tools that amplify a person and tools that harvest one. Almost everything now being built with AI is technically capable of either. Which one it becomes is decided early, in the social and technical architecture, by the people holding the reins. That’s why I do this work.
And I’ll say the quiet part plainly: I think that whatever we build together deserves to be built with heart, integrity, and love infused in every line of code and every corner of the interface. If that sentence makes you feel more at home here rather than less, we will probably get along well.
“Harlan cares deeply about people, and that shows in everything he does. He’s passionate about building excellent tech… and innovating towards the future.”
Harlan T Wood
I’ve been building software for twenty-five years and never wanted to stop. I ran my own consultancy for six years before I ever held a job title at someone else’s company, then spent two decades leading and building inside engineering teams in San Francisco and New York — developer, architect, director, CTO — including blockchain prototypes for Fortune 100 companies and six years as principal architect at a financial-services platform.
I learned lean software from the Poppendiecks, the Core Protocols from Jim and Michele McCarthy, and information design from Edward Tufte. I’ve been told my most useful quality is that I can sit in a room with engineers and with the people who aren’t engineers, and leave with everyone understanding each other.
I have also spent about as many years in meditation, personal growth, and transformational work as I have in software, and I have never kept the two in separate boxes. I listen with empathy. I resonate with the mission, and with a founder’s commitment to build with integrity and deliver polished software.
“He has an impeccable personal integrity that generates good will, good results, and good luck. I value any opportunity to work with Harlan and I recommend him highly.”
Tell me who you are and what you’re making. If there’s a fit, we’ll find a time. No pitch, just a conversation about whether I can help — and lots of good advice to send you on your way, regardless.
