Hello there, my name is

Mike Heijmans

I build intelligent hardware from silicon to cloud: embedded systems, platform engineering, and applied AI.

I build intelligent hardware from silicon to cloud. I spent roughly twenty years on platform and infrastructure engineering at scale: Kubernetes, CI/CD, observability, and globally distributed teams. Now I also design the hardware underneath, from PCBs in Altium and ESP32 firmware in C to the cloud platform that manages the fleet. Applied AI is the through-line: multi-agent coding pipelines, MCP servers that let agents drive real applications, and the observability to measure whether any of it actually works. Most engineers build one end of that stack; the interesting problems live in owning both.

Silicon to Cloud

As CTO of Rising Orchards I build IgorBox, a show-control platform for haunted attractions and live venues: cloud-connected controllers that drive the lights, motors, and DMX rigs on the floor. This is production engineering, not a weekend project, and I build every layer of it:

Firmware: ESP32 in C on ESP-IDF, with OTA updates, WiFi provisioning, and multicast coordination that keeps a venue full of controllers in sync.
Hardware: schematic and PCB design in Altium, 4-layer boards taken from prototype through production.
Interfacing: isolated RS-485 and DMX, motor drivers, and LED drivers. Unglamorous work, but it's what makes hardware safe and reliable enough to run a show with a paying audience in the room.
Regulatory: FCC and ISED certification, the difference between a working prototype and a product you can legally ship.
Cloud: a Next.js platform on Vercel that authors and sequences shows, then deploys them to a fleet of networked controllers in the field.

Roughly twenty years of platform and infrastructure engineering behind the software; the hardware is where I get to build the whole thing from the copper up.

Applied AI

AI isn't a spectator sport for me. I use it, build with it, and instrument it every day:

I run multi-agent development workflows where coding agents work across multiple repositories, negotiate API contracts with each other, and hand off work through Linear.
I build MCP servers that connect agents to running applications, turning AI from a code generator into a developer that can test its own work.
I instrument my AI usage with OpenTelemetry and make workflow decisions from real token and cost data, not vibes.
I write about all of it: context engineering, how attention bias shapes agent behavior, and what actually works in production.

Leadership

As a manager and director, I believe people are a company's greatest asset. When you unlock each individual's potential, everyone benefits. I've built and led remote distributed teams of engineers, designers, and product managers spanning multiple time zones, and I'm passionate about creating a culture of collaboration, trust, and respect. I'm a strong advocate for diversity and inclusion, and I strive to create an environment where everyone can contribute and grow.

Beyond the Keyboard

In my free time, I enjoy working on open source projects, learning new technologies, flying airplanes (sometimes upside-down), scuba diving into caves, and making music. I'm passionate about creating and building new things, and I love to share my knowledge and experience with others. I'm always looking for new opportunities to learn and grow, and I'm excited to see what the future holds.

Recent Writing

Multi-Repo AI Coding: Move the Context to Linear

7/10/2026

12 min read

A while back I wrote about using requirements markdown files as context for AI coding sessions, and then about why loading that context fresh at the top of a session works so well. That workflow is still the foundation of how I code with AI. But recently I hit its ceiling.

The ceiling is called "a project that spans multiple repos." An API repo, a frontend repo, a firmware repo, an infra repo, and multiple Claude Code sessions working across them, sometimes at the same time. The requirements-file-in-the-repo approach starts to crack here, because the question becomes: which repo? So I moved the project-level context out of the repos entirely and into Linear. It has been one of the best workflow changes I've made this year.

I'm Getting 20x My Money's Worth From Claude Code (Here's the Data)

6/29/2026

4 min read

I recently wired up AI Center (Coralogix's AI observability product) for coding agents on my local development hosts. (yes it's plural... I use multiple computers to develop and run coding agents). Its original intention was to just see token usage and efficiency of coding sessions. Plus it was a fun project to play with AI Center. It turns out there was more to the data than I realized and a genuinely useful data point I hadn't considered. As a Claude Code Max subscriber, I found out I'm getting WAY MORE than my money's worth on that $200 a month cost.

I Built a Vercel Logdrain Bridge to Coralogix

3/28/2026

4 min read

I don't think it's a secret that I love Coralogix and I also love NextJS. I run quite a few NextJS apps on Vercel and I wanted to send my backend logs to Coralogix so I can correlate trace data to log data. It turns out that Vercel has "Logdrains" which can send your logs to a webhook endpoint. The problem is the format is not Coralogix compatible so I whipped up a little bridge application to transform the data and send it to Coralogix. Here's how it works and how you can use it yourself.