Skip to content

/about · spec sheet

Michael Barry.

I ran restaurants for twelve years and studied music business before that. Then I taught myself to code and shipped my first production software in 2021. Since then, I’ve built data-heavy web platforms for the water industry: 50+ water resource clients, hundreds of production features. Regulated, unglamorous, high-stakes software. The good kind.

Timeline

  • before restaurants

    A music business degree, then a year at a record label

    B.A. Music Business at SUNY Oneonta in 2009, then a year at A&M/Octone in New York, whose roster at the time included Maroon 5, K'naan, Flyleaf and Hollywood Undead. By 2010 restaurants were next.

  • before software

    Twelve-plus years of restaurant general management

    Landry's restaurants at Disney World, where T-Rex was the second-highest-grossing restaurant in the country. Opening Del Frisco's Grille in Westwood. Barcelona Wine Bar in Denver. Then the Denver Milk Market, sixteen venues, through COVID. P&L ownership and team leadership throughout, which is where the commercial half of my current job came from.

  • before going pro

    Under a year teaching myself web development

    Udemy courses, books, YouTube, documentation, Stack Overflow, and building things until they worked. Nearly all my free time, with serving shifts on the side to make ends meet.

  • 2021 – present

    Full-Stack Engineer & Team Lead, LRE Water (remote)

    Web platforms for agencies, districts and utilities across the US: 50+ client organizations, from permitting and water accounting to flood telemetry. I own applications from the first conversation through to maintenance years later, plus the half of the job that isn't typing: watching for bids, writing the parts of a proposal that describe the software, helping set scope and budget.

  • 2022 – present

    Michael Barry Web Development, LLC founded

    It began as subcontract work: UI features on PineSpire's electrification dashboard, through another developer's LLC that held the account. Then I took on clients of my own (a gym, an asset-recovery company, a neighborhood nonprofit) and built and shipped their work end to end.

  • 2022 – present

    The AI-native shift, in stages

    AI has been in the work since roughly 2022, and what changed across those years is how much of it I handed over. The change that mattered was not a better model: it was writing specs and review loops instead of prompting my way through a task. Over a year now since my last traditionally typed line of code.

  • 2024 – present

    Leading the core development team

    Six of us in Water Information Technology core development, three of them my direct reports. Core is the data-heavy water resource web tools; a separate team carries the rest of the division's work, including the complex modeling.

How I operate

Why water
The work has a cause. It isn’t a landing page for a bank, and it isn’t one small feature inside somebody else’s large product. Consulting adds volume and variance: enough repetition across clients to build for reuse, enough difference that nothing is ever a template. I was the only developer on the UI side for most of those years, which is how one job came to cover maps, data management, authentication and authorization, visualizations, public tools, forms, schema and database code, security, backend APIs, widgets, admin tools, reports, exports, and the support calls afterwards. A lot of autonomy, and a lot of surface area. I doubt a narrower job would have taught me as much in five years.
AI, at company scale
I was invited to the first AI pilot at a company of around four hundred: ChatGPT and Codex, Claude and Claude Code, GitHub Copilot, Vercel’s AI tools, evaluated against each other. Narrower pilots followed as things got adopted, a Claude Code team pilot among them, and I’m now part of the initiative building company-wide internal tools. The rest of it is presentations, to the company and to my own team, and mentoring. Making this workflow safe for one developer is a different problem from making it safe for an organization. I’d rather be in the room for the second one.
How I work
I build the systems that write the code: documentation packages, agent prompts, acceptance criteria, review loops. The typing is delegated; the judgment, and the learning, are not. The whole pipeline, with the real artifacts, is on /systems.
How I know it works
The obvious question about all of that, and the one I’d ask too. The answer is that the gates live in the repository, not in my good intentions: lint, ast-grep drift rules that turn an architecture violation into a build failure, types, unit tests, the build, then end-to-end tests behind all of it. Agents can’t merge past a red check any more than I can. The actual workflow file is on /systems.
Working with clients
The job starts before the repository exists: working directly with project managers and the organizations I build for to turn a loosely defined operational need into a scoped application with an architecture and a budget. I watch for bid opportunities and write the parts of a proposal that describe the software. The first conversation and the maintenance years later are the same job.
Working with a team
I lead a development team, and I ran restaurant teams for twelve years before that. Both taught me the same trick: put the lesson where it cannot be missed. A drift rule whose error message names the fix teaches more than a paragraph in a README, and a guidance package that ships in the same release train as the code cannot quietly go stale.
Seeing the code
Most of the day job is private by nature, because the code belongs to the firm and to the clients rather than to me. The activity is still visible on github.com/MB13534 ↗ even where the repositories are not, and if you want to read the actual code, ask and I’ll walk you through it.

What I’m looking for

The other half of a portfolio.

Everything above is what I have done. This is what I want next, so we can both find out quickly whether this is a fit. The list is short on purpose.

A hard problem
Work that is genuinely difficult: real constraints, real consequences, a domain with rules that punish guessing. Regulated water data turned out to be a gift that way. The problem should be harder than the tooling, because the tooling now moves very fast.
Ownership, end to end
The first conversation, the architecture, the ship, and the maintenance three years later. I have done the job that way for five years and I would rather not give up any part of it. Give me a queue of tickets and you get closed tickets. Give me an application and you get an owner.
Autonomy, with a compass
Trust me with the how, be specific about the why. Total autonomy is mostly abandonment with better branding, so that is not the ask: I want a clear goal, room to choose the approach, and someone senior enough to argue with when I choose wrong.
Room to grow
Somewhere I am not the most experienced person in the room, at a scale I have not worked at yet, next to people who are ahead of me on something. Five years in, I would like the next five to be steeper.
People to build with
Twelve years of restaurants and five of software taught me the same lesson twice: the work is a group activity. I lead a development team now and I like it. Direct reports or not, I want colleagues who review my thinking and not only my pull requests.
People to build for
The pattern in my work is not an accident: 50+ water resource clients and the public behind them, a neighborhood gym, a neighborhood nonprofit, a neighborhood fitness platform. Software with someone specific on the other end of it. Community is not a perk on that list, it is why the details matter.
AI-native, at organization scale
Making this workflow safe for one developer is the problem I have already solved. Making it safe for a team, and then for a company, is a different problem, and it is the one I want next. If you are working that out, I would like to be in the room.

Stack

Frontend

Charting

Mapping

Backend

Data

Auth

Tooling / testing

Deploy / ops

Elsewhere

Other things I’ve done.

Twelve years of restaurants, from Disney World to a sixteen-venue food hall. A music degree and a year at a record label. A fruit grove, nine chickens, four rooms rebuilt. A bicycle to the top of a 14,000-foot mountain. A book that AI wrote and illustrated and I printed for my daughters. And a lot more where all of that came from.