GitHub repos for developers

Robert C. Martin's Clean Code principles adapted for JavaScript, with short bad and good examples covering naming, functions, objects, classes, SOLID, testing, concurrency and error handling. A quick way for a junior JavaScript developer to start writing code that survives review. Treat the rules as guidelines, as the author himself says, rather than dogma.

The REST API design guidelines Microsoft uses internally, covering URL structure, HTTP methods and status codes, error formats, pagination, filtering, versioning and long-running operations, with companion guides for Azure and Microsoft Graph. They are useful as a concrete, opinionated checklist when designing your first real API. Some rules reflect Microsoft platform needs, so adopt the reasoning rather than every detail.

John Crickett's Coding Challenges have you build real tools from scratch: wc, a JSON parser, a Redis server, a load balancer, a DNS resolver, a shell, a git client. This repo indexes community solutions to each challenge in many languages. Attempt the challenge yourself first, then read two or three solutions in your language to compare how other people structured the same problem.

The source of the first edition of Addy Osmani's book, free to read online. It explains the classic patterns (module, revealing module, observer, mediator, facade, factory, singleton) in JavaScript, with the reasoning for when each helps and when it is overkill. The repo has not changed since 2021 and parts predate modern ES modules and classes; the second edition (O'Reilly, 2023) and patterns.dev cover the modern JavaScript and React versions.

A long directory of engineering blogs from companies, individuals and products, from Netflix and Uber to one person newsletters. The most useful part is the included OPML file: import it into a feed reader and prune it down to the twenty blogs you actually read. It was last updated in 2024, so some blogs have moved or gone quiet.

One link hub spanning data structures and algorithms, low level design and high level design: pattern write-ups for dynamic programming, sliding window and binary search, video lectures from MIT and Abdul Bari, practice sites, LLD questions with solutions, and database design reading. It is a directory rather than original material, and was last updated in December 2024, so expect the occasional dead link.

A curated list of engineering blog posts, papers and talks organised by problem rather than by company: rate limiting, caching, distributed consensus, message queues and their antipatterns, load balancing, time series databases, distributed transactions, and real systems such as Google Docs and a chess engine. Most links go to the teams who actually built the system, which makes it a good place to see how a concept holds up in production.

A reading list of the original papers behind modern infrastructure, grouped by area: storage (the Google File System, Haystack, HDFS), analytics (Monarch, Gorilla), cluster scheduling (Borg, YARN), stream processing (MillWheel, Dataflow, Flink), logs and pub/sub (Kafka, LogDevice) and graph processing (Pregel). Reading the source beats a fourth hand summary. There is no commentary and the links are LinkedIn short URLs, so pair a dense paper with a write-up if you get stuck.

Hundreds of short, focused snippets and articles, mostly JavaScript with CSS, Python and more, each explaining one technique in about the time it takes to read. Reading many small, correct examples is one of the quickest ways to pick up a language's idioms: array methods, closures, DOM helpers, date handling. Actively maintained, and 30secondsofcode.org is the easier way to browse and search it.

A community written, deliberately exhaustive answer to "what happens when you type google.com into your browser and press enter?" It starts at the keyboard interrupt and works through DNS, ARP, sockets, the TLS handshake, HTTP, HTML parsing, CSS and rendering. Reading it once is a fast way to find the layers you use every day but could not explain. It has not been updated since 2024, so newer pieces such as HTTP/3 and QUIC are not covered.

A compact reference for the data structures and algorithms that come up again and again: linked lists, trees, tries, heaps, graphs, sorting, greedy approaches, bitmasks and runtime analysis, each explained in a few plain sentences with its time complexity. It also links online judges and has a directory of worked solutions. The top of the README promotes a paid newsletter; the reference sections below it are the free part worth reading.

The free, open companion to Arslan Ahmad's Grokking the System Design Interview course. It teaches the pattern first approach: understand the building blocks (caching, sharding, replication, queues) well enough that any design question becomes a matter of combining them, instead of memorising answers. Includes a seven step framework for structuring a design discussion, a question catalogue, cheat sheets and study roadmaps, all readable in the browser without an account.

Anthropic's free, official notebook courses: API fundamentals, an interactive prompt engineering tutorial, real world prompting, prompt evaluations and tool use. Work through them in order and you will have built each technique yourself rather than read about it.

Kelsey Hightower's step-by-step guide to building a cluster with no installer: certificates, etcd, the control plane, worker nodes and pod networking, each configured by hand. You would never run a cluster this way, and that is the point; afterwards nothing in a managed cluster is a black box.

A free, self-paced course from DataTalks.Club covering experiment tracking with MLflow, orchestration, batch and online deployment, monitoring and best practices like testing and CI/CD, with weekly homework and a peer reviewed capstone. It makes you ship a complete pipeline rather than read about one.

Adds vector similarity search to Postgres, so embeddings live next to the rows they describe and you keep transactions, joins and backups you already understand. For most first RAG projects this is all the vector database you need, and the README is a clear explanation of HNSW, IVFFlat and hybrid search.

A large, actively maintained collection of notebooks where each one implements a single retrieval technique, from basic chunking through hybrid search, reranking, query rewriting and evaluation. Useful because you can compare techniques on the same problem and keep only the ones that actually improve your results.

GitHub's own free exercise that has you create a branch, commit, open a pull request and merge it in a real repository, in under an hour. The quickest way to go from knowing Git commands to having done the collaboration workflow once.

A curated set of articles, talks and books about the part of the job that is not writing code: scope, influence, communication and ownership. Useful when you are technically solid, keep getting good reviews, and still cannot work out what is missing for the next level.

Thousands of questions and exercises across Linux, networking, Docker, Kubernetes, CI/CD, cloud, SQL and Terraform, with answers. Good for finding the holes in what you think you know, whether or not there is an interview coming.

The index of every curated awesome list, covering nearly every language, framework and topic in software. When you are starting in an unfamiliar area, finding its awesome list first will usually save you a week of searching.

A directory of free APIs sorted by category, with authentication and CORS support noted for each. The fastest cure for not knowing what to build: pick an API that interests you and make something small with real data instead of another to do list.

A very large, actively maintained index of free books, courses and cheat sheets, organised by language and subject and available in many languages. Check here before buying anything. A surprising number of standard texts are legally free from their authors.

Hundreds of algorithms implemented readably, with a repository for each major language: Python, Java, C++, JavaScript, Go, C# and more. Useful in two ways: as a reference when a textbook description has not clicked, and as a first open source contribution, because adding a missing algorithm is a genuinely approachable pull request.