System Design.

Mission: System Design, via real big-tech architectures

Why

A working engineer across frontend and backend who has shipped plenty of CRUD but never run into real scale. The goal is tech-lead scope: be the person in the room who can write a design doc, review someone else's design, and argue a trade-off credibly — not by reciting patterns, but by knowing what Facebook, Discord, Notion and Uber actually did and why they chose that price. Curiosity is a co-driver: the real architectures are interesting on their own terms.

Success looks like

  • Reads a proposed design and names the failure it will hit first, with the mechanism, not a vibe.
  • States a trade-off in one sentence in the shape big-tech papers use: "we trade X to reduce Y."
  • Can retell 8–10 real architectures (FB memcache, Discord messages, Notion sharding, ...) from memory, including the numbers.
  • Writes a design doc for a real change at work that a senior reviewer signs off without a rewrite.
  • Knows which of these problems the systems at work do not have yet — and can say so out loud in a planning meeting.

Constraints

  • No deadline. Optimise for storage strength (long-term retention) over cramming.
  • Starting point: solid API + DB CRUD. No prior work with sharding, replication lag, queues, or multi-region.
  • Every lesson must be anchored to a real, documented company system — no invented toy examples as the centrepiece.
  • Lessons stay short: one win, completable in a sitting.

Out of scope (for now)

  • Coding interview / DSA prep.
  • Hands-on ops: running Kubernetes, Terraform, cloud console work.
  • ML systems design and AI infrastructure.