About Systems Unpacked
Systems Unpacked is for working engineers who want deeper understanding of the systems beneath their software. We connect small, inspectable examples to performance, reliability, and AI workloads. The published learning material is free.
Make a prediction. Inspect the mechanism. Explain the result.
Each lesson introduces the vocabulary, works through an example, and gives you a changed scenario to reason about. Explanations, code, evidence, and hints stay together. Start at the beginning or jump to the question you need.
Explanations you can inspect. Results you can question.
Lessons state their assumptions, link to sources, and distinguish measured behaviour from simplified models. Worked answers explain both what the evidence shows and what it leaves open. Every activity carries one of these labels:
- Model — a small deterministic teaching model. Its steps are not CPU cycles or real timings.
- Recorded evidence — saved output from a stated Linux environment. Native experiments include recorded results and explanations; nothing native runs live on this site yet.
- Constructed case — a deliberately built scenario for reasoning, with explained answers.
All core reading and reasoning can be completed without a terminal or installation.
Review status
The two foundation sections are authored and author-checked. Independent technical review has not happened yet; each lesson shows its own status. If something is wrong or unclear, tell us.
Author
Karel Moravec writes the lessons and builds the site. He is a software engineer who spent years working on backend systems and wanted a resource that connected how computers actually work to the questions that come up in practice — performance, memory, scheduling, AI serving — with small models and real evidence instead of hand-waving.