About

Distributed systems.
Useful software.

I’m Rich Majewski, a Senior Distributed Systems Engineer. My current work focuses heavily on high-performance computing (HPC) and AI in heterogeneous hyperscale data centers.

My work today

At ThisWay Global, I work on Amalgamy.ai, a platform for orchestrating HPC and AI workloads. My contribution spans workload planning and resource allocation, integration between components, and the practical work of making a platform deployable and operable.

The interesting problems rarely stay inside one component. A planning decision has consequences for the allocator. An installation decision has consequences for running workloads. An interface that looks convenient today can make a different deployment unnecessarily difficult later.

My job is to work through those connections, turn requirements into explicit behavior, and help make the pieces agree. That involves architecture and implementation, but also testing assumptions, explaining tradeoffs, and revising a design when the evidence says it needs to change.

The decisions between the boxes

A diagram can make a distributed system look deceptively tidy. The arrows are where a lot of the work lives.

Take node maintenance. Before an upgrade, something needs to know whether the node is busy, stop new work arriving, let existing work finish, and know when it is safe to return the node to service. I want that lifecycle expressed clearly at the platform boundary, rather than teaching every component the details of one scheduler.

That is the kind of question I spend time on: where does a responsibility belong, what must remain true when things fail, and how do we demonstrate that the implementation does what we intended? The goal is a system people can reason about, not one that only its original author can rescue.

Three decades in infrastructure

I’ve worked through three decades of infrastructure evolution, from operations and private cloud to platform engineering and distributed systems. That includes production environments at multi-thousand-node scale and the automation needed to make them manageable.

I’ve built private-cloud provisioning workflows, worked on infrastructure migrations, and owned the reliability of customer-facing device fleets. The recurring problem is often the work people do to compensate for gaps between systems. I would rather fix the gap than make someone keep bridging it by hand.

That’s the part of engineering I keep coming back to: understanding why something is difficult, then changing the system so it doesn’t have to stay that way.

It started with Zork

I got into computers because I wanted to build a text adventure. I had a D&D story in my head and wanted a computer to let someone else explore it.

I learned by typing programs out of magazines and library books, debugging my typos, and trying to make pieces of different programs work together. The game never quite happened. The learning did.

I’m self-taught in computing. I still tend to start with a picture of how something could work, then pull it apart until I understand what it would take to build it. Sometimes that becomes useful software. Sometimes I learn why the idea was wrong.

What I’m building

My public projects include Plainwright, a Markdown editor, and DwarFSctl, alongside other KDE desktop applications, archive systems, terminal libraries, and build tooling. Some are everyday tools; others are places to work through a harder engineering question.

I’m especially interested in making useful engineering accessible. Someone mounting a compressed image shouldn’t need to understand filesystems. Someone checking memory use should get an explanation, not just a pile of counters.

I also explore AI-assisted development and how to divide the work between human judgment, generative tools, and checks that can actually verify a result.

See the projects
A Nile monitor lizard
A different kind of company.

Away from the keyboard

And then there’s the monitor.

This is Pixie, my Nile monitor. I’ve spent years earning his trust, learning what makes him feel safe, and remembering that trust doesn’t switch off instinct.

I’m also drawn to fantasy, mythology, and the creatures people call monsters. Looking frightening and being malevolent are very different things.

Get in touch

For a project question, its GitHub repository is usually the best place to start. You can also find me on GitHub and LinkedIn, or send me an email.

Looking for a résumé?

The short version of my background is here. If you need the full professional history, request my résumé.