Complexity & designFree
No Silver Bullet: Essence and Accidents of Software Engineering
Fred Brooks, 198618 min
The courses senior engineers wish they had
Engineering mastery for the AI era.
AI writes more of the code every month. Knowing how systems fit together, fail and scale is what lets you direct it and review what comes back. Every lesson is drawn, narrated and quizzed.
Live · tixdrop, Friday 10:00
- Arriving
- 900/s
- Finishing
- 800/s
- Waiting
- 1,200 · 1.5s
[01] Watch · Press play, no sign-up
- Animated lessons
- 84
- Papers, close-read
- 18
- Hours of video
- 32
- Free to watch
- 36
Learners from
[02] Figures
Every idea gets drawn.
Every figure comes from a real lesson and uses its numbers. Four of them run live, so go ahead and break them.
Hover to playTap to playFig 01
900 requests arrive, 800 finish, 6,000 wait in line
Why distributed? The single machine ceiling
LiveClick a replicaTap a replicaFig 04
W=2, R=2 on three replicas: every read meets the write
Dynamo: the cart store that always says yes
Hover to playTap to playFig 05
4,000 engineers wrote the file, #4,001 has to read it
Why Go, why now: a 2007 language built for the AI era
[03] Courses
Courses that go deep.
Each course takes one topic all the way, from first principles to the systems you run. Every one starts free, so you can judge the teaching before you pay.
25 lessons · 5 hours · Free to start
Distributed Systems in a Day
40,000 fans, 5,000 seats, one drop at ten o'clock. Follow it from one server to a hundred times the load: why every piece exists, and how it breaks.
- 01Why distributed? The single machine ceiling11 min
- 02TCP: the first arrow14 min
- 03TLS: a secret in public13 min
- 04Service discovery: finding payments mid-drop11 min
33 lessons · ~14 hours · Free to start
Go Foundations
The most complete Go foundations course out there. 33 lessons and ~14 hours, from hello world to a tested, concurrent HTTP service on Go 1.27. You build feedstack, a real feed aggregator, one lesson at a time, and the course is updated with every Go release.
- 01Why Go, why now: a 2007 language built for the AI era12 min
- 02The toolchain tour: installing Go with mise, run, build, fmt, vet8 min
- 03Your first program: compile it, run it, see what happened13 min
- 04Values, variables, constants, and zero values15 min
17 lessons · 4h 34m · Free course
Data Structures & Algorithms: Foundations
The 80% every engineer needs, from first principles. Complexity, the memory hierarchy, the core structures, and the algorithms built on each.
- 01Complexity: big-O, amortized, and the cost model that matters14 min
- 02The memory hierarchy: why arrays are fast14 min
- 03Arrays and dynamic arrays from scratch13 min
- 04Strings: immutability, builders, and the hidden O(n)13 min
9 lessons · ~3.3 hours · Free to start
Design a Message Queue
Build the queue behind a shop's order emails, from a list in memory on one machine to a replicated broker like SQS.
- 01What a queue promises23 min
- 02The log on disk21 min
- 03fsync and the loss window20 min
- 04Receiving without losing a message23 min
Or follow a track.
Courses in a suggested order for one goal. Every course stands alone, so start wherever you are.
- 3 of 12 liveBackend engineerBuild and run the server side of a product: HTTP, APIs, databases, auth, caching, queues, and what changes when one server is not enough.
- 2 of 9 liveSystem design interviewLearn the building blocks, then work through the classic designs the way an interview runs them: requirements, estimates, a first design, then deep dives.
- 0 of 8 liveAI engineerShip product features built on language models: how inference works, retrieval, agents and tools, evals, and running it all in production.
- 0 of 7 liveCoding with agentsGet real work done with AI coding agents, from a first session in your repo to agents that run on their own in CI.
- 1 of 5 liveCoding interviewThe data structures and algorithm patterns coding interviews test, each taught with the real system that uses it.
- Courses still to come show as open steps, and fill in as they ship.See all tracks
[04] Papers
The papers behind the work.
Classic and recent papers every engineer should know, each explained in an animated video and a close-reading write-up. Some are free to watch.
[05] From learners
I've been working as a backend engineer for a while, and most of my system design understanding came from scattered blogs, docs, and bits of papers. Karnstack helped bring that into a much clearer structure.
Anamika
Lead Software Engineer at Headout
The course hits that sweet spot between depth and breadth of the topics involved in distributed systems, making it a great way to learn/revisit the fundamentals. Its interactive simulators and quizzes keep things fun and engaging while leveling up the architectural intuition from first principles.
Bala Vijaya Krishna Saladi
Software Engineer at Amazon
I'm a frontend engineer who'd been faking my way through system design conversations for years. This course is the first thing that actually closed the gap. The communication module alone, tcp, tls, service discovery cleared up more than every half-read blog post combined. Now I can sit in a design review and actually follow what the backend folks are arguing about.
Karandeep Singh
Senior Frontend Engineer at Appbrew
[06] What’s new
Shipped lately.
New lessons land every week, and the rest of the site keeps changing too. This is the log.
[07] Next
More courses are next. You pick.
36 ideas, one topic each, in the same animated format. The most voted get made first.
- 01
Claude Code in an Afternoon
60 minInstall it, learn the agent loop, then steer it with plan mode and permission modes on a real repo.
- 02
Context Engineering for Coding Agents
30 minCLAUDE.md, memory, compaction and checkpoints: what the agent sees, and how to keep it sharp in long sessions.
- 03
Skills and Hooks
45 minTurn your repeated prompts into self-loading skills and your rules into hooks the agent cannot skip.
- 04
MCP from Zero
40 minHow the Model Context Protocol works on the wire, then build a server that gives an agent your own tools and data.
- 05
Subagents and Parallel Work
50 minWorktrees, subagents and background sessions: run several agents on one codebase without them colliding.
[08] Questions
Questions, answered.
Format, access, refunds. If yours isn’t here, email mail@karnstack.com.
Platform
Purchasing
Help
Start with a free lesson. Stay for the course.
The data structures course is free, and every other course opens with free lessons. Plans and prices are on the pricing page.