// matt lantz · ai systems · application architecture

I build AI systems that actually ship.

Software engineer and application architect for 15+ years. For the last five I've been putting ML and agentic systems into production: scoring film and TV data, running social media operations, and turning municipal records into summaries residents can actually read.

Book a 30-min call See the work

Open to my next full-time role · taking select consulting work through Grafite

Portrait of Matt Lantz
Matt Lantzai systems · architecturehuman in the loop
agent.tracelive

Workflows

$ agent run civic-digest --meeting council
planfetch → extract → summarize → checkok
toolfetch_agenda(source="")ok
toolextract_items(agenda.pdf)ok
modelsummarize(items, reader="resident")ok
evalfaithful · cited · plain languagepass
toolqueue_review(editor)ok
✓ digest ready for human review
Real workflows. Client names withheld.

15+

years shipping production software

5

years putting ML & AI into production

3

years building agentic systems

2

monolith-to-microservices migrations, one in progress

Built software at

BlackBerry · Pfizer · Amadeus Hospitality

Laravel framework contributor

// selected work

Most of it runs behind a login.
Here's what it does.

Client names stay private. The problems, the architecture and the results don't have to.

Case 01 · Entertainment data

Read the case study

A scoring engine for film & TV

Problem

Film and TV data arrives from many sources, in many shapes, and rarely agrees. Teams needed one score they could trust to rank and compare titles.

What I built

Ingestion and normalization pipelines, ML scoring models, and the services that put those scores in front of the product.

Outcome

325,000 · bi-weekly · Improved content ranking and discovery

  • ML scoring
  • Data pipelines
  • Entity matching
  • Scoring API
  • PHP, Laravel, MySQL
Case 02 · Social mediaAgents that run a social calendarDrafting, scheduling and reporting across channels.Case 03 · Municipal governmentCouncil meetings, summarized for residentsCollects agendas and records, pulls out what matters, writes it in plain language.Case 04 · Personal opsAn agent-run back officeAgents built into the tools I use to run my own projects and clients.

// architecture

Monolith to microservices. I'm on my second.

Splitting a large monolith into services while the business keeps running is one of the hardest jobs in software, and few engineers get to do it even once. I've done it for a multimillion-dollar business, and I'm in the middle of doing it again for another.

  • Service boundaries that match how the business actually works
  • Data and traffic moved over in stages, with old and new running side by side
  • Teams keep shipping features while the platform changes underneath them
Read the case study
migrations.log1 running
  1. 01monolith → microservices✓ shipped

    multimillion-dollar business · hospitality · 2

  2. 02monolith → microservicesin progress

    multimillion-dollar business · manufacturing · now

Two large migrations. Client names withheld.

// what i do

Where I add the most value.

Whether you're filling a role or scoping a project, this is the work I'm best at.

  • 01

    Agentic workflows

    Agents that plan, call your tools and hand off to people at the right moment, not a chatbot bolted onto a page.

  • 02

    Monolith to microservices

    Breaking up a large monolith without stopping the business. I've done it once and I'm partway through my second.

  • 03

    ML scoring & data pipelines

    Turning messy, multi-source data into scores, rankings and classifications a product can rely on.

  • 04

    AI in existing products

    Adding LLM features to the Laravel, PHP, Java and Vue codebases you already run, without a rewrite.

  • 05

    Evals & reliability

    Checks on model output, guardrails, fallbacks and cost controls, so AI features hold up in production.

  • 06

    Architecture & technical leadership

    Application and cloud architecture on AWS or Azure, code review and mentoring from someone who has shipped software for 15+ years.

// in public

The parts you can see.

Open source contributions, products I've launched, and places I've written and talked about building software.

Open source

Products I've built

Writing & talks

Writing

All writing

Experiments

All experiments

// next step

Hiring, or have an AI feature or migration that needs to ship?

I'm looking for my next full-time role and taking on select consulting work. A 30-minute call is the fastest way to find out if I can help.

Book a 30-min call LinkedIn