How This Course Works + Your Study Plan

Here's an honest number: this curriculum is 344 pieces — around 300 lessons, a dozen hands-on codelabs, 37 checkpoints, and five capstone passes, across five courses. Stated flat like that, it's a little terrifying.

So this lesson does two things, both practical. First, it tells you exactly what you'll be doing here — the four kinds of thing a system-design lesson can be, and why the course is shaped the way it is (there's real learning science under it, not decoration). Second, it hands you a route: three honest study plans for three kinds of person, and a 30-second chooser to find yours.

Because 344 items isn't a wall. Done right, it's about 45 minutes a day — and the next lesson's diagnostic will even tell some of you to skip most of it.

Scope: this is the map of how to travel. The very next lesson, Diagnostic: Where Should You Start?, picks your exact entry point; One System, Five Passes: Meet Your Capstone explains the project thread. Here we cover the routes and the rhythm.

Every item in this course is one of four types, and it helps to know them on sight:

• Lessons (most of it) — one concept, taught from a concrete problem down to the trade-off it forces. Every lesson ends with two things on purpose: a "The Trade-off" section (when to use it, when NOT) and a "🧪 Try It Yourself" beat. That last one isn't filler — hold that thought, it's the whole reason the course works.
• Codelabs — hands-on, copy-paste-runnable. Actually spin up nginx, produce to Kafka, break a Postgres replica and watch it fail over. Reading about a load balancer and running one are different kinds of knowing.
• Checkpoints — a short decision quiz at the end of each section. Not recall trivia — scenarios ("which cache strategy for this workload, and why?").
• Capstone passes — five of them, one after each course. You re-design the same system — a food-delivery marketplace — five times, each time with the new tools you've earned. One system, five passes, beginner to staff.

You've already done three lessons and met the marketplace. You know the rhythm; now here's why it's built this way.

That Try-It beat and those checkpoints look like extras you could skip to "save time." Skipping them is the single most common way smart people waste months. Here's the science, and it's unusually clear-cut.

In 2013, a landmark review (Dunlosky and colleagues) graded ten popular study techniques on how much they actually improve learning. Highlighting, re-reading, summarizing — the things most people do — landed in low utility. Only two techniques earned high utility:

1. Practice testing — retrieving an answer from your own head, not re-reading it. Roediger and Karpicke showed the effect is large: students who tested themselves crushed students who re-read, even though re-reading felt more productive. That's what every checkpoint and Try-It is — a tiny retrieval rep.
2. Distributed practice — spacing learning across days instead of cramming it into one. Same total hours, dramatically better retention.

So the course is deliberately built to be recalled, not re-read: short lessons, a retrieval beat in each, a checkpoint per section, and routes measured in days rather than weekends. The design is doing half the work for you — but only if you don't skip the parts that feel like effort. The effort is the learning. That's not a metaphor; it's the finding.

There is no single right way through this — it depends on your goal, your clock, and where you're starting (yes, The Trade-off Mindset applies to studying too). Three honest routes cover almost everyone:

🏃 The 2-Week Interview Sprint (~2h/day). Interview in days? Don't start at lesson 1. Go straight to C5 — Interview Mastery (the delivery method, the 12 problem patterns, the technology deep-dives, the L4/L5/L6 calibration) and drill the 20 practice problems. When a pattern references a concept you're shaky on, follow the cross-link, learn just that, come back. Fast and gap-aware — you'll know what you don't know, which is most of the battle.

🚶 The 6-Week Solid Track (~1h/day). A month or two to prepare properly. Walk C1 → C2 → C5 in full, skim the heavier C3/C4 deep-dives, and do one capstone pass a week. This is the sweet spot for most working engineers with a real interview on the calendar.

🧗 The 3-Month Mastery Route (~45min/day). No deadline, or you want this to stick for your career, not just an onsite. Every course in order, C1 through C5. Every codelab, every checkpoint, all five capstone passes. And here's the quiet superpower: because of the spacing effect, 45 minutes a day for three months genuinely beats a two-week cram at the same total hours. Slow is fast.

Three questions — goal, daily time, current level — and the route that fits you lights up. It also flags the honest edge cases: the beginner with an interview on Tuesday, the senior who can skip the on-ramp.

Try this first: answer as your real situation, not your aspirational one. "90 minutes a day" only counts if it's true. The route below the recommended one is always a valid, simpler fallback.

The routes trade the same thing every design decision trades: depth against time.

The sprint buys speed and pays in shallow spots — you'll pass with strong fundamentals and a few "I know the shape but not the internals" answers. Mastery buys durable, transferable understanding and pays in patience. Neither is wrong; they fit different constraints, exactly like every architecture in this course.

The one genuinely bad move is the mismatch: cramming the 3-month route into a weekend (the spacing effect guarantees you'll forget most of it) or grinding all of C1's hardware lessons with an interview 48 hours away. Match the route to the constraint. And write your route down with a flip condition, like everything else: "6-week track — but if the interview moves up, I drop to the sprint."

Two minutes, and it makes the next three months real:

1. Pick your route in the chooser above — honestly.
2. Put it on a calendar. Not "I'll study system design" — an actual recurring block: "7:00–7:45am, System Design Mastery." Spaced practice only works if the spacing is scheduled.
3. Write one flip condition: "[route] — unless [what would change it]."

Check: if your plan is "I'll do a few hours this weekend," you've just designed the one approach the learning science says fails. Re-read Why the Checkpoints Exist — then trade the weekend binge for 45 minutes tomorrow, and the day after, and the day after that.

• "More hours = more learning." → Same hours, spaced, beats massed — decisively. 45 min/day for a month buries a weekend cram. It's one of the most replicated findings in cognitive science.
• "I have to do all 344 items, in order." → No. The routes exist precisely so you don't. An engineer interviewing next week who starts at lesson 1 is making a scheduling error, not a diligence virtue.
• "Checkpoints and Try-Its are optional filler." → They're the highest-utility technique known (practice testing). Skipping them to "save time" is skipping the actual learning to keep the feeling of it.
• "Reading the lesson means I learned it." → Recognition isn't recall. You learned it when you can retrieve it cold — which is what the Try-It checks and the interview will too.
• "I should build a detailed study schedule before starting." → Pick a preset, block the time, start lesson 1 today. Over-planning is procrastination with a to-do list.

• The course is four item types — lessons (ending in a trade-off + Try-It), codelabs, checkpoints, and five capstone passes over one running system.
• It's built to be recalled, not re-read: the Try-Its and checkpoints are practice testing, and the day-a-time routes are distributed practice — the only two techniques rated high-utility by the research. The effort is the learning.
• Three honest routes: 2-week sprint (C5 + problems), 6-week track (C1→C2→C5), 3-month mastery (everything, in order). Match the route to your goal and clock.
• Spacing beats cramming at equal hours — so schedule the time, don't binge it.
• 344 items is really ~45 minutes a day. Write your route down with a flip condition and start.
• Next — Diagnostic: Where Should You Start?: 15 questions that pick your exact entry point into everything you just mapped.