papers every engineer should know
15 classic and modern papers that shaped how we build software, walked through one at a time. close-read, no summaries.
15 papers|5 modules|5 hours
what you’ll learn
- read foundational systems papers without getting stuck on the notation
- extract the actual idea from a paper in 30 minutes instead of an afternoon
- spot the difference between a paper that aged well and one that didn't
curriculum
01
module one
foundations & philosophy
~45 min3 papers01
No Silver Bullet: Essence and Accident in Software EngineeringFred Brooks, 1986the cleanest articulation of why software stays hard.
free8m02
Out of the Tar PitBen Moseley, Peter Marks, 2006accidental complexity, and a functional-relational escape hatch.
21m03
End-to-End Arguments in System DesignSaltzer, Reed, Clark, 1984one paper that changes how you think about where to put logic.
11m02
module two
systems & operating systems
~45 min3 papers04
The UNIX Time-Sharing SystemRitchie, Thompson, 1974how seven people designed a system we still live inside.
18m05
Reflections on Trusting TrustKen Thompson, 1984turing award lecture; the compiler-backdoor idea is pure fun.
free20m06
On the Criteria to Be Used in Decomposing Systems into ModulesDavid Parnas, 1972the origin of information hiding; every senior review comment descends from here.
24m03
module three
data & algorithms
~50 min3 papers07
A Relational Model of Data for Large Shared Data BanksEdgar F. Codd, 1970why your database looks the way it does.
19m08
Space/Time Trade-offs in Hash Coding with Allowable ErrorsBurton Bloom, 1970bloom filters. a delightful "wait, that works?" paper.
19m09
HyperLogLog: Analysis of a Near-Optimal Cardinality Estimation AlgorithmFlajolet, Fusy, Gandouet, Meunier, 2007counting billions of distinct items with kilobytes of RAM.
22m04
module four
scale & production
~50 min3 papers10
The Tail at ScaleJeff Dean, Luiz Barroso, 2013the p99 problem, explained by the people who lived it.
22m11
Dapper, a Large-Scale Distributed Systems Tracing InfrastructureSigelman et al., 2010the paper every observability vendor is still copying.
21m12
The Log-Structured Merge-TreeO'Neil, Cheng, Gawlick, O'Neil, 1996why rocksdb, cassandra, and every modern kv store look the way they do.
19m05
module five
modern AI & ML systems
~55 min3 papers13
Attention Is All You NeedVaswani et al., 2017the 8-page paper that ate the world.
22m14
Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksLewis et al., 2020the pattern half the industry is building on now.
18m15
Efficient Memory Management for LLM Serving with PagedAttentionKwon et al. (vLLM), 2023a systems paper wearing an ml hat. the bridge between both halves of this course.
21mfrequently asked
- do i need to have read the papers first?
- no. each lesson close-reads the paper with you from the first page. a copy open alongside helps but isn't required.
- what if i can't afford a plan?
- email mail@karnstack.com from a student address (.edu, .ac.in, equivalent) and we'll work something out.
- is there a refund?
- yes. 7 days, no questions, mail@karnstack.com.
how this course is made
the curriculum is curated by karnstack and reviewed by senior engineers in the industry before it ships. narration is an ai voice (elevenlabs) reading human-written, human-reviewed scripts. read how courses are made.