Tech IndustryDec 13, 2023
Airbus GroupTotalLost

Chance of clearing E6 MLE at Meta.

I have my interview at meta in 3 months for MLE role. I am ML research Lead, with ML PhD (average not top) and 8 years of experience at the same company. I am working a team lead on research project. So, I don’t have any experience in system design or code in production. My project are cross-functional and organisational, where I lead 8-10 people, setup the strategy, roadmap, the technologies bricks to work on … I have less and less time to code. I don’t consider myself a good dev either. But I work with a nice team and help each other. I have 3 months to prepare the interview and I am taking these 3 months off to prepare. I will be dedicating 50h a week. I have to learn from scratch leecode as I have never used it before. And also, systems design as I have never designed any system and I don’t now anything about it. Have to learn all the concepts and do some mocks interviews. In total it is: 600h of preparation. Is that enough to clean the interview? And also I am worried about my no-existing experience in designing systems. And I get a big chance to clean the interview, what are the chances to survive as E6 at meta? If I get an E5 role (with chance too), I am also worried about my coding skills as I am not expert and I don’t write nice code (it’s a spaghetti researcher code 😂😂😂). Please comment if 3 months (50h a week) is enough to clear the interviews (no knowledge or experience on SD). I am really scared.

Poll
31 Participants
Select only one answer
Flagged by the community.
Airbus Group TotalLost OP Dec 13, 2023

Wow thanks a lot for you help. Is it enough for the interview 3 months of preparation ?

Microsoft Mk0l2 Dec 13, 2023

Hey i just want to say that I’ve benefitted a lot from your channel and it certainly helped landing my E5 at Meta.

Amazon OIsp28 Dec 13, 2023

I don’t know that much about Meta or MLE, but if it’s anything like an SDE role, you might struggle to clear the phone screen for a junior role based on your description. Anything is possible, I guess, but it took me years of incremental progress to really understand what it takes to get an engineer to give you good interview feedback which is very different from being able to solve LeetCode problems, which is hard enough on its own. Anyways, I would suggest investing in LeetCode Premium and doing their guided learnings on interview patterns you need to know (such as DFS, BFS, binary search, heaps, linked lists, etc.). Neetcode.io is another really good website/YouTube channel. Try to utilize mock interviews with real people as much as possible. Look at sites like interviewing.io for that after you gain basic competency with leetcode easy-medium problems. Some may disagree with this ethically, but it definitely helps to do real interviews with other companies you care about less before the one you really want, although I wouldn’t put it at the every end as that might add too much psychological pressure. After that, if you get hired, it’s all about utilizing your company’s internal search tools to see code examples and documentation. You also need to master your IDE and integrate it with your company’s tools and it’ll do half the work (of writing code, there’s lot of other stuff it can’t help with). Good luck!

Airbus Group TotalLost OP Dec 13, 2023

Thanks a lot for your comment. I am not bad at coding, I am coding for the last 20 years. I just said that maybe my code is not well optimised.

Meta djwnnq Dec 13, 2023

If you’ve never designed a system before it will be a real struggle to operate at E6

Airbus Group TotalLost OP Dec 13, 2023

I have designed system but not at scale, like ML systems using database on aws, api. I have been building RAG systems with LLM for the last 2 years, serving lot of users but not millions.

New
qFrY02 Dec 24, 2023

600h hours prep sounds like over-preparing.. For the record I recently got rejected for E6 for the second time. probably prepared around 200hr and I think it was probably too much. nailed coding and behavioral but apparently did bad on ML design. I disagree. The interesting part is that my 4/6 interviewers were E5s. I know because I asked them directly. My key comments are: 1. It doesn’t matter how much time you prepare, be ready for at “least” 50% interviewers bias. 2. Even if you LC for 300hr and you miss an edge case it’s most likely you will miss it in the next 300hr. So prepare more on optimality not on quantity. Applies the same for all other rounds.

Airbus Group NewLife24 Dec 25, 2023

Thanks a lot for your advices. Can you share your preparation ressources? The ML system design question you had ?

Capital One zlxmn Jan 5

wtf the only comment that seemed to be helpful in this thread was flagged by the community?!

Meta W0lverinee May 17

can someone please share what was mentioned in the first comment? Looks like it was the most useful one, and it got flagged by the community.