Lab

Things I’m testing, learning, and trying to understand.

The lab is intentionally unfinished. It is where ideas can be useful before they become projects, and where experiments are allowed to fail without pretending they were products.

Active terrain

The areas I keep returning to because they expose interesting systems underneath the surface.

Infrastructure

Self-hosted systems

Linux, Docker, Kubernetes/MicroK8s, reverse proxies, databases, local services, observability and recovery behavior.

recurring interest
AI systems

Models as components

LLMs connected to APIs, tools, local logic, memory, observability and external systems—not just used as chat interfaces.

recurring interest
Networking

Failure paths

DNS, routing, segmentation, service reachability and the difference between symptoms and actual boundaries.

Software

Useful glue

Python, JavaScript, APIs, automation and small tools that connect larger systems together.

Hardware

Physical systems

Sensors, compute, peripherals, electronics and the point where software stops being abstract.

How experiments graduate

A small rule set to keep curiosity from turning every test into an endless platform.

Question

What am I actually trying to learn?

If the question is vague, the experiment will expand indefinitely.

Proof

What result would count as success?

A working API call, a measured behavior, a reproduced failure, a deployed service, or another concrete artifact.

Checkpoint

Stop when the question is answered

Then decide whether to stabilize it, archive it, or deliberately promote it into a larger project.

“Build enough of the system to expose the mechanism. Don’t build an empire just to answer one question.”