Do you remember this part in the book "Designin Data-Intensive Applications by Martin Kleppmanns" about how twitter approached home timeline feature?

What they basically did, that instead of querying DB for all the tweets you follow - they are doing push approach, so you are basically reading from queue that is getting inserted when your followee pushes new tweet. One thing that for celebrities they left "query" approach.

I found the conference where it is explained with data, all path, problems, etc, very interesting:
https://www.infoq.com/presentations/Twitter-Timeline-Scalability/

LETS GO

motherfuckers, prev time 3 years ago they didn't have it I remember

Today we are streaming, usual schedule 20:00 Prague time. Leetcode, System Design + discussing AWS screening 😁😁😁

Started:

https://www.youtube.com/watch?v=bRofQ83SLb4&ab_channel=andreyka26_se
https://www.twitch.tv/andreyka26_

https://www.youtube.com/watch?v=G3cpafmjZ1c&ab_channel=andreyka26_se

got connection dropped, created new stream

Lets hack Pentagon

LET"S Go.

But if we are serious, this is exactly how many days you need before the interview. 1-2 leetcode problems per day, cover all the patterns (sliding window, array, 2 pointers, dp, dfs, bfs, graphs, etc) and you are kind of ready.

https://leetcode.com/problems/rotate-image/?envType=study-plan-v2&envId=top-interview-150

literally my feeling, after I solved my stupid way, and looked at the solution)

https://www.interviewbit.com/

Interesting thing, recruiter from FB recommended this thing to my friend instead of leetcode:

has anybody tried it?

Well well well. It seems like doesn’t matter how high you are. And I hardly can imagine how these top managers are looking for a job, applying and interviewing. Usually, what I saw, such guys are 20+ years at the same company, so it is not like there is huge flow in and out there.

At least for engineering - it is easily justifiable: be able to write code, be able to design stuff, have relevant exp

One more:

Stream in 2-3 hours

Today we are talking about important stuff about the graphs. Union Find, dfs, bfs.
On top of that I wanted to review the MIT lecture about BST node deletion and compare with our solution, as I think MIT lecture was a bit not complete.

As the last thing - we are reviewing solutions for URL shortener on Youtube and compare with ours, and why they better or worse.

If we have time left - we will finish URL shortener impl + load tests