Editorially curated clusters. Each topic collects the posts that belong together, in reading order.
Algorithm write-ups organised by pattern: greedy, dynamic programming, graphs and union-find, sliding windows, linked lists and binary trees, heaps and binary search. Every post starts from the brute-force solution, walks step by step to the best time and space bounds, and ends with runnable code and a complexity analysis.
What it actually takes to put a language model into production: how multi-agent systems split responsibility, how sub-agents talk to each other, how task boards and rate gates prevent cascading failures, and where the trust boundary sits when text from a web page tries to become an instruction.
The full chain from interface design to production operations: Django and DRF practice, rate limiting and circuit breaking, idempotency, caching and message queues, plus the domain, TLS, CDN and host choices that come up when a service actually ships.
This topic cares about one thing: how bytes actually get written down and read back. Page splits and back-to-root writes in a B+ tree, the write-ahead ordering of a WAL, the version chain behind MVCC, and the trade-offs between LSM and B-trees — all taken down to specific pages and fields.
Complete teardowns of open-source projects: a browser with an in-process CDP implementation, a zero-runtime Rust agent, a 13 MB Windows SSH client, a single-file Markdown reader, a local-first personal assistant, and a 30k-row interview-experience pipeline. The point is never "what I built" but which trade-offs were made and what broke along the way.