leetcode-patterns VS grok_sdi_educative

Today, after updating my personal website, I shifted focus to a crucial aspect of tech interviews: data structures and algorithms. Starting with arrays, I've committed to tackling one topic per day, solving 3ish related LeetCode problems to reinforce my understanding. 📚 (This website is super useful - sorting leetcodes by patterns)

He says things that are pro-labor. Then he does things like oppose minimum wage increases, support anti-union actions, appoint judges who will oppose unions or labor, appoint two very anti-labor assholes to the NLRB, and outright just not pay for labor. Trump is not "pro labor", he's just a liar.

Quality over quantity! I solved hundreds of them while preparing for big tech interviews and I can say with confidence after certain number of questions we have diminishing returns. A good base of questions with patterns clearly explained is all you really need. As of now there are many lists online Blind 75 (https://www.teamblind.com/post/New-Year-Gift---Curated-List-of-Top-75-LeetCode-Questions-to-Save-Your-Time-OaM1orEU), Prashad's leetcode patterns (https://seanprashad.com/leetcode-patterns/), curated list of problems from Elements of Programming.

Grinding leetcode is inefficient. What you should be doing is familiarizing yourself with the common patterns you might expect to see in an interview. Look at the blind 75 and https://seanprashad.com/leetcode-patterns/.

Initially, you don’t need to solve any of the problems from scratch. Look up the problem on YouTube and someone will walk you through it. This will build your intuition of when to reach for a heap or for a DP array or when to do BFS, etc. If you don’t know these, then watch another video explaining the concepts. These videos are often 10-15 minutes a day so with a 30 minute time commitment you potentially can get through 3 a day, getting you through the complete blind 75 (more than enough) in less than a month or 1 of each of SP’s 22 patterns in a couple of weeks.

The great thing is you don’t need dedicated time for this approach, you can often start a video while tackling laundry or doing some dishes.

Then, start putting these into practice but spend no more than 10-15 minutes on the problem. If you can’t solve it, go watch the video again. There are so many times where you can have the right approach but make a stupid mistake that will cause you to flounder and you can pick up a better way of doing it. Eventually you will be solving these in 10-15 minutes and the time commitment will have remained at a minimum.

After this, find a new job that is only 40 hours a week and voila you’ve just opened up 10-20 hours for personal projects.

I used this list of patterns for my preparation, but you can use another; it's just an example of what it usually looks like. 😊

Sean Prashad has a good list that maps the type of problem to a particular pattern.

Sean Prashad: 212+54 problems

When I started grinding I had already taken a DS course so I was at least aware of hashmaps, arrays, trees, etc. but even leetcode easies we’re still a struggle and the ones I could do I wasn’t solving them optimally. Having that basic knowledge though allowed me to start struggling through the book I mentioned. EPI is harder than CTCI but it’s language specific and I already knew I was going to use python. I completely skipped bit manipulation and tries but I ended up getting a trie question in one of my interviews and that was by far the hardest question I’ve ever gotten, probably because i skipped tries altogether lol. I haven’t tried grokking. I started doing EPI by hand so that I was forced to think through solutions before writing them but if you feel like that’s unnecessary for you just type it but it def helped me since in interviews you have to discuss before coding. The DP section of epi was really hard so I didn’t complete it. From there I just started struggling through LC. After epi I started with easies but i didn’t stay there for long since epi is mostly mediums and some hards anyways. The biggest thing is to learn from every problem, even the ones you don’t finish. Truly take the time to understand a solution even if that means 45+ minutes of digging into it. The same patterns come up again and again so you’ll have to dig into them eventually. Identifying the patterns from the way questions are phrased come from quantity I feel like so doing a lot of LC problems is the way to go. My pattern recognition was on point after about 100 problems in total maybe. For LC i did 95% of all easies and mediums in [this]. Earlier I said I skipped DP in epi but def do dp leetcode problems. You don’t have to master it but def become familiar. If by the end of everything you’re able to knock out 70+ medium level questions then you should probably be good for internship interviews. Hope this helps and also if you take the blind 75 route NeetCode on YouTube has done all of them I think and he’s really good with explaining the problems. Take the time to understand each data structure individually before even attempting problems. I think EPI does a good job of that or you can just hop on YouTube. Good luck Glum Choice, I have faith that you will prevail 🏅

LC refers to Leetcode problems (LC = LeetCode). While the Leetcode website itself is not the only place to practice, "Leetcode problems" refers to the general concept of programming interview problems involving data structures and algorithms.

I am also new at leetcode. But from the advices I gathered and from my own experience, I will suggest to first learn all data structures. Once you understand, then start with the easy ones and gradually learn the patterns. Sean Prashad as a good list of the pattern you can find via this link: https://seanprashad.com/leetcode-patterns/

Grokking Systems: github.com/sharanyaa/grok_sdi_educative

grokking the system design

Originally I took a udemy course which explained all the building blocks, like Network, caching, indexing, db etc. Once you know those you can start reading how these combine to make a system (https://github.com/sharanyaa/grok_sdi_educative gives all standard examples). I would also recommend to read blogs on large companies like Netflix, Linkedin etc to get an idea of real world examples alse.

I went through Facebook and Google interviews (as well as others) back in early 2019 and eventually accepted an offer from FB for an M1 role (was previously a senior manager at a smaller company). At FB I had a coding slot, a couple of system design slots, and I think three behavioral. At Google I didn't get any coding, but it was because of a higher level role (L7) that I was interviewing for (and did not get).

For context, I hadn't coded professionally for ~8 years at that point, so part of the prep was getting back into that. Also, I never took any algo classes in college 15-20 years earlier, so that was another challenge. I actively avoided big tech companies for years due to difficult interviews and low confidence that I would pass, not to mention a general attitude of why the hell should I spend all this time preparing for something that does not reflect the day-to-day reality (and I still hold this view strongly). But as I learned more about compensation opportunities, I decided to bite the bullet and go for it. I quit my job and spent 3-4 months preparing mostly full-time, while also lining up interviews. Pro-tip here: if you get Google interviews, start them way in advance of all others. Their process is insanely long and after 3 months of interviewing, re-interviewing, getting downleveled, changing hiring managers, etc. I dropped out of the process because I was already sitting on two offers. For comparison, Facebook took 2 weeks from starting to schedule interviews to offer.

For coding, I picked up Python - unless you know something else really well, it's IMO the best language for interviews (never used it professionally, but had not great experience from a few years back interviewing in my language of choice of C#). I got a few books to learn and refresh on the data structures and algorithms: Algorithm Design Manual (good reference, but I didn't end up having time to really go through the material there), Cracking the Coding Interview (skimmed, basic to intermediate stuff here) and Elements of Programming Interviews in Python (skimmed, intermediate to advanced stuff here). I got lucky and recalled the general approach to my FB coding problem from the EPI book. Now for the most important part - create an account on leetcode.com and start practicing. I would recommend focusing on easy and medium levels. I imagine managers would generally get medium-level questions, engineers are expected to solve medium and advanced. My approach was to identify a list of topics, then for every topic read ADM, then CtCI, then EPI, then solve a few easy and medium LC problems, rinse and repeat. I've seen folks recommend that solving ~200 LCs is enough to get comfortable. I didn't have as much time as I would've liked and did about 60. The general idea is to be able to pattern match (e.g. this type of problem needs a hash table, this is a graph problem, etc), some people focus on memorization - I think that's a terrible approach, but most companies do end up asking questions from LC (except Google apparently). With limited time I was flying through topics towards the end, would not recommend. I did pay for Leetcode Premium for a few months - lets you focus on the types of problems that specific companies are asking, was worth it for me.

Coding aside, I actually spent most of my time preparing for system design interviews. Found that my knowledge of building large systems was not nearly sufficient for interviews. My approach here was to first purchase this course: https://www.educative.io/collection/5668639101419520/5649050... and read through the basics section. It really is basic, so I was supplementing by diving deep into each topic on my own. It's still a pretty useful resource for getting an idea of what breadth I should be targeting, along with https://github.com/donnemartin/system-design-primer. Then I read Designing Data-Intensive Applications (amazing book, but incredibly dense), and finally went through all the concrete design problems on the Grokking course as well as watched a bunch of relevant Youtube videos. Felt a lot more comfortable towards the end and I think did alright on those slots. I think the concept of consistent hashing rings came up in literally every single system design interview, so be super familiar with it and various implementations. My interview experience with these was that the really senior folks were more into depth, and the less senior ones focused on breadth. I'd suggest going for both, e.g. the Grokking course and System Design Primer have decent coverage of breadth of topics, but are absolutely not enough for depth. Maybe you could skate by on breadth alone (I'd say it's more valuable than depth), but why take the chance. Some examples of going deep I encountered were talking about different kinds of load balancers and balancing mechanisms, different ways of traversing a consistent hash ring, different failure modes and workarounds in distributed systems. I originally wasn't going to memorize common latency numbers, but it came up during Google's prep session (they offered one with an interviewer, to give feedback) when I embarrassingly struggled to convert 1 Gbps to 125 MB/s.

Of course most rounds were non-technical, so I prepared for those with as many questions I could come up with in advance and recalling various scenarios in detail - super important not to assume that these will just come to you during an interview. I also put together a 65-page summary doc of all the stuff I was reviewing, most of it system-design related, that I went over frequently to refresh myself. Finally, the Blind app and levels.fyi were instrumental in opening my eyes on the crazy compensation at those companies and I was able to do a better job negotiating with this info.

I was fortunate that I had the ability to take the time off to focus on interview prep, and I would 100% not have passed the interviews without it. My cynicism about the whole process only grew, as exactly nothing from my prep was relevant to the day-to-day job at Facebook. ¯\_(ツ)_/¯

https://github.com/sharanyaa/grok_sdi_educative (pirated though)

https://github.com/donnemartin/system-design-primer https://github.com/sharanyaa/grok_sdi_educative/blob/master/grok_system_design_interview.pdf

I started here: https://github.com/sharanyaa/grok_sdi_educative. It’s just a print from the course with that name on Educative (but from 2 years ago).