When it comes to solving challenging problems, a person’s color, background, dating preferences, or accent matter none to me.
It’s all about vibes, skill, intellect, and the ability to solve hard problems.
Unstructured, synchronous logging is a scalability tax.
https://skilledcoder.medium.com/how-to-log-1m-events-sec-without-slowing-down-your-system-1a64cfd536b9
Obsessed with this series!
A really good read for distributed data folks.
How and Why Netflix Built a Real-Time Distributed Graph: Part 2 — Building a Scalable Storage Layer https://netflixtechblog.medium.com/how-and-why-netflix-built-a-real-time-distributed-graph-part-2-building-a-scalable-storage-layer-ff4a8dbd3d1f
Yesterday, finished 4 series of technical grilling (quite fun).
1. Matrix mutation and traversal.
2. Systems deep dive for building a recommendations system.
3. Building metrics for a system, creating the calculations then writing the window functions for them.
4. Deep dive into troubleshooting DB issues at scale.
Today, back-to-back 🏋️♂️training sessions today then I’m logging off to enjoy my weekend.
Hi, I’m Ashley. I collect beautiful system diagrams the way others collect art.
But seriously, if you’re learning these concepts. The goal is always low latency high throughput.
Latency - how fast a message is sent
Throughput - how many messages per second can be sent
lol I think this is funny! You can internally build every service in this list at a smaller scale and specific to your usecase. These are all generalized abstractions of primitives.
Been noticing an uptick in matrix traversal questions lately which totally tracks with the AI age we’re in.
Everything from embeddings to attention maps lives in matrix form. So mastering traversal, rotation, and pathfinding isn’t just LeetCode grind it’s literally foundational linear algebra for the AI era.