Serving a billion web requests with boring code - llimllib notes
When I worked as a contractor to the US government at ad hoc, I was fortunate enough to get the opportunity to design large parts of a relaunch of medicare plan compare, the US government site through which hundreds of thousands of medicare recipients purchase their health care plans each year. We released after about a year of development, and were able to help millions of people find and purchase health care - if you're in the US, you are pretty likely to know somebody who used this system. Though the US health care system is incredibly awful in many respects, I'm very proud of what the team I worked with was able to build in a short time frame and under a lot of constraints. The team of people that I worked with - managers, designers, engineers and business analysts - were excellent from top to bottom. I've never before experienced such a dedicated, collaborative, trusting team, and I learned a tremendous amount from them. I especially want to share this story because I want to show
Serving a billion web requests with boring code - llimllib notes notes.billmill.org / blog / 2024 / 06 / Serving a billion web requests with boring code last updated: Jun 24, 2024 High level Boring über alles The boring bits Why React? golang The innovation tokens Modular backend gRPC Strict backwards compatibility Faceted search Database creation ETL models testing Local database for development Miscellaneous tooling Logging Documentation Runtime integrations And more When I worked as a contractor to the US government at ad hoc , I was fortunate enough to get the opportunity to design large p
related reading
- Building a web search engine from scratch in two months with 3 billion neural embeddingsblog.wilsonl.in
- Everything I know about good system designseangoedecke.com
- Patterns for Building LLM-based Systems & Productseugeneyan.com
- How my website worksbrianlovin.com
- How I side project | thesephist.comthesephist.com
- Low Latency and Model Training at Modalrhea24.github.io
- What We’ve Learned From A Year of Building with LLMs – Applied LLMsapplied-llms.org
- Let’s Handle 1 Million Requests per Second, It’s Scarier Than You Think!youtube.com
- Scaling long-running autonomous coding · Cursorcursor.com
- "We ran out of columns" — the best, worst codebase — Jimmy Millerjimmyhmiller.github.io
- Extensible Software in the age of LLMs | Jeremy Morrelljeremymorrell.dev
- There's no reason for software to be slow anymoredanluu.com