Lesson 1 · Free · 2 min
$ echo $((400000 / 86400))
Watch lesson 1 freeWhy design starts with questions · 2 min
System design for senior engineers, AI systems included
Architecture · 26 lessons · 59 min · AdvancedLearn to design and defend production systems the way senior engineers do: requirements, estimates, APIs, trade-offs said out loud. You extend one product, Lumen, from a simple app to a retrieval and agent system, then run a capstone design that includes a failure and its diagnosis.
What you'll be able to do
- Run a design from requirements and honest estimates to an API, a data model and a high-level diagram
- Choose between caches, queues, replicas, search and sharding, and state the trade-off you accepted
- Make retries, timeouts and overload safe with idempotency, budgets and load shedding
- Design a retrieval pipeline, a vector search layer and an inference path against a latency and cost budget
- Give an agent tools with least-privilege permissions and put evals into the design itself
- Diagnose a production incident from the signals a design should expose
How the lessons work
Lessons that move
Narrated, animated lessons with real terminals and code on screen.
Pause and ask
Stuck? Ask out loud or type. The answer plays as a short video made for your question.
Check yourself
Short quizzes inside the lessons show whether it stuck.
Course content
26 lessons · 59 min · 2 lessons are freeUnit 1 · A method you can repeat4 lessons · 9 min
01Why design starts with questions2 minFree02Estimating with honest assumptions2 minFree03The API and the data model first2 min
04The high-level design, said out loud2 min
Unit 2 · Building blocks and when they earn their place6 lessons · 13 min
05Caching that stays correct2 min
06CDNs and HTTP caching2 min
07The database and its replicas2 min
08Queues and asynchronous work2 min
09Search is a derived copy2 min
10Rate limiting2 min
Unit 3 · Scale and reliability4 lessons · 9 min
11Sharding and the partition key2 min
12Retries and idempotency2 min
13Back-pressure and load shedding2 min
14Failure domains and graceful degradation2 min
Unit 4 · Designing the AI parts6 lessons · 14 min
15The model as an unreliable component2 min
16A retrieval pipeline, end to end2 min
17Vector search choices3 min
18Inference under a latency and cost budget3 min
19An agent with tools and permissions2 min
20Evals are part of the design2 min
Unit 5 · Full designs3 lessons · 8 min
21Design: a notification service2 min
22Design: permission-aware answers3 min
23Design: a multi-tenant AI gateway3 min
Unit 6 · Capstone: Lumen goes live3 lessons · 7 min
24Capstone: the scenario and requirements2 min
25Capstone: the design2 min
26Capstone: the incident and its diagnosis2 min