available for workalex@axxe · Netherlands · agents · automation
AXXE
alex@axxe ~ % whoami

I build agents
and ship the busywork

I'm Alex, an AI engineer with a decade-plus of shipping software. I set up agents, automate the work nobody wants to do, write agentic code, and help teams get good at building with AI. Based in the Netherlands; Dutch & English.

40+
agents & workflows shipped
9
teams leveled up
100%
code you own
An ops agent that runs the weekly close on its own.Featured case
NORTHWIND·Agents

An ops agent that runs the weekly close on its own.

6h → 4mprocess time
// Selected work

Things I've shipped.

Agents, automations, and the times I helped a team learn to build it themselves. Filter by the kind of work.

Showing 3 of 3 · plus more in the archive
// Writing

Ideas & concepts.

What I'm learning building agents and helping teams adopt them. Thinking in public.

All writing →
Essay

Your first agent should be boring.

The temptation is to start with the flashiest workflow. Start with the one nobody will miss if it breaks.

May 2, 2026
6 min
Note

Evals are the actual product.

A demo proves an agent can work once. An eval suite proves it keeps working. Guess which one ships.

Apr 18, 2026
4 min
Essay

Agentic coding changes what 'senior' means.

Less typing, more judgment. The bottleneck moves from writing code to deciding what is worth building.

Mar 29, 2026
8 min
Playbook

A checklist before you let an agent write to prod.

Guardrails, dry-runs, and the human-in-the-loop gates I wire into every agent that can touch real systems.

Mar 10, 2026
5 min
Alex, AI engineer
Alex
AI engineer · Netherlands · Dutch / English
// about

A bit about me.

I've spent a decade-plus shipping software: leading backend teams, architecting cloud-native systems, and delivering AI platforms from enterprise apps to chatbots. These days I point all of that at agentic systems and automation: LLM agents, RAG over a vector store, and MCP tool integrations. What separates a demo from something you can actually ship is the boring layer around it: evals, guardrails, and human-in-the-loop gates. That's the part I care about most.

I care about two boring things that make systems last: security and maintainability. OWASP habits, SOLID code, tests and evals you can actually trust. And I like leaving teams better than I found them: half the value is the system, the other half is your people knowing how to build the next one.

Stack I reach for
LLM agentsLangGraphMCPEvalsRAG + pgvectorClaude / OpenAI / AzurePythonTypeScript
Currently
available for new work

Building agents that earn their keep, and helping a couple of teams get fluent in agentic coding. Open to projects from the Netherlands and remote.

The only way to do great work is to love what you do.

// Trusted by teams shipping with AI
NORTHWINDACMEHELIOGRAPHBRACKETVANTA LABS

We replaced a 6-hour weekly process with an automation that runs in four minutes. The code is in our repo and our own engineers maintain it.

Dana Reyes · VP Operations, Northwind
6h → 4m
weekly process time
$220k
annual cost removed
11 days
scope to production
// Get in touch

Got something worth automating?

Whether it's an agent to build, a workflow to kill, or a team to level up, tell me what you're wrestling with and I'll tell you straight if I can help. I read every message myself.

Prefer email? Reach me directly at alex@axxe.tech