Hi folks, Just got contacted by a FB recruiter for a Data Engineer role. They mentioned I should brush up on Python and SQL but didn't give a ton of specifics on what topics to cover. Was curious to learn from folks who might have gone through the process and work at Facebook as DE - 1. how did you prep for the interviews 2. What areas to focus on most 3. Are behavioral interviews a thing at FB? About me: Started as a Business Analyst with Amazon and used to work a lot on SQL (still do). slowly picked up python using in-house trainings and helping out fellow Data Scientists/Engineers with their data tasks. Moved over to a Business Intelligence engineer role couple years back and started working with native AWS tools, redshift admin, tableau EC2 servers, etc. Currently doing LC easy questions on Arrays and taking 20+ min on those using brute force approaches (which isn't the ideal way to do it). Figuring out better solutions usually involves me looking at discussions/solutions and implementing those logic back on the original question. Any help/guidance on what topics to cover and in what depth will help me make the best use of my limited prep time (interview scheduled end of March) #data #dataanalytics
Interested to pair up and prepare
Dm'd you
Interested to pair up and prepare
Im in the same boat, interested to pair up
Interview Format: Tech1: Product Sense questions, 2 SQL questions (2nd ques was on ETl), 1 Python problem Tech2: Product Sense questions, 1 SQL question, 1 Python problem DM: Product Sense questions, designing a DM followed by 4 SQL questions Data Engineer Analytics Generalist role, - No spark/map reduce questions - SQL/Python are based on the product sense ques (and not Leetcode problems as is) - Python probs are focussed on dictionary/list manipulation (and not algo based really) - ETL question was to write sql code to load data from staging to target and not a design question really Resources for Onsite: SQL/Python: practice on Leetcode and https://pgexercises.com/questions/aggregates/ DM: practice data models for these products: uber, netflix, dropbox, messenger, imdb, airbnb https://www.vertabelo.com/blog/example_models/ https://allbigdatathings.blogspot.com/2019/04/uber-data-model.html Product Sense: read about user engagement metrics (DAU/MAU/WAU), funnel metrics and read some cases from the following website https://stellarpeers.com/blog/ Blind/RoofTop Slushie posts - https://www.rooftopslushie.com/request/Facebook%20data%20engineer%20interview-4375 - https://www.rooftopslushie.com/request/Facebook%20Data%20Engineer%20Onsite%20Interview-410 - https://www.rooftopslushie.com/request/Facebook%20Data%20Engineer%20Analytics%20Generalist-7794 - https://www.rooftopslushie.com/request/Facebook%20DE%20Onsite%20Virtual%20Interview-5606 - https://www.rooftopslushie.com/request/Facebook%20Data%20Engineer%20Interview-1253 - https://www.rooftopslushie.com/request/Facebook%20Data%20Engineer%20interview-5383 - https://www.rooftopslushie.com/request/Facebook%20data%20engineer%20analytics%20onsite%20interview-8046 --------------FB PhoneScreen Format: 5 SQL and 5 Python For phonescreen I prepared fully on Leetcode (probs under Facebook tag) SQL: Hard/Medium Python: Easy (tags: list, string, hashtable) previous questions from Glassdoor: SQL: 1. What percent of all products in the grocery chain's catalog are both low fat and recyclable? 2. What are the top five (ranked in decreasing order) single-channel media types that correspond to the most money the grocery chain had spent on its promotional campaigns? 3. % Of sales that had a valid promotion, the VP of marketing wants to know what % of transactions occur on either the very first day or the very last day of a promotion campaign. 4. What brands have an average price above $3 and contain at least 2 different products? 5. To improve sales, the marketing department runs various types of promotions. The marketing manager would like to analyze the effectiveness of these promotion campaigns. In particular, what percent of our sales transactions had a valid promotion applied? 6. We want to run a new promotion for our most successful category of products (we call these categories “product classes”). Can you find out what are the top 3 selling product classes by total sales? 7. We are considering running a promo across brands. We want to target customers who have bought products from two specific brands. Can you find out which customers have bought products from both the “Fort West" and the "Golden" brands? 8. what %age of products have both non fat and trans fat. 9. find top 5 sales products having promotions 10. what %age of sales happened on first and last day of the promotion 11. Which product had the highest sales with promotions and sales ( basically a where clause on 2 flags) 12. Manager want to analyze the how the promotions on certain products are performing.find how the the percent of promoted sales? 13. get the top3 product class_id by the total sales. 14. Percentage increase in revenue compared to promoted and non-promoted products. 15. Products classes that has the highest transactions 16. Count of Customers who bought 2 items type (A,B) Python: 1. Complete a function that returns the number of times a given character occurs in the given string For example: - input string = "mississippi" - char = "s" - output : 4 2. Fill in the blanks Given an array containing None values fill in the None values with most recent non None value in the array For example: - input array: [1,None,2,3,None,None,5,None] - output array: [1,1,2,3,3,3,5,5] Ensure you take care of case input[None] which means None object. 3. Complete a function that returns a list containing all the mismatched words (case sensitive) between two given input strings For example: - string 1 : "Firstly this is the first string" - string 2 : "Next is the second string" - output : ['Firstly', 'this', 'first', 'Next', 'second'] 4. Find common words in 2 sentences 5. Given two sentences, construct an array that has the words that appear in one sentence and not the other. 6. Given an array of integers, we would like to determine whether the array is monotonic (non-decreasing/non-increasing) or not. Examples: // 1 2 5 5 8 // true // 9 4 4 2 2 // true // 1 4 6 3 // false //1 1 1 1 1 1 // true 7. Valid ip address, edge case to remember is if there are alphanumeric characters. 8. Count distinct words in a sentence 9. Count the number of times a word appear in a sentence using a Hash Map 10. Return tuples of a list, matching each item to another item 11. Count the number of times a substring appear in a string 12. Calculate the average word length. For the given set of words return the average word length. str.split() 13. want you to write me a simple spell checking engine. The query language is a very simple regular expression-like language, with one special character: . (the dot character), which means EXACTLY ONE character (it can be any character). So, for example, 'c.t' would match 'cat' as the dot matches any character. There may be any number of dot characters in the query (or none). Your spell checker will have to be optimized for speed, so you will have to write it in the required way. There would be a one-time setUp() function that does any pre-processing you require, and then there will be an isMatch() function that should run as fast as possible, utilizing that pre-processing. There are some examples below, feel free to ask for clarification. Word List: [cat, bat, rat, drat, dart, drab] Queries: cat -> true c.t -> true .at -> true ..t -> true d..t -> true dr.. -> true ... -> true .... -> true ..... -> false h.t -> false c. -> false */ // write a function // Struct setup(List<String> list_of_words) // Do whatever processing you want here // with reasonable efficiency. // Return whatever data structures you want. // This function will only run once // write a function // bool isMatch(Struct struct, String query) // Returns whether the query is a match in the // dictionary (True/False)
This awesome man, thank you so much
Are you done with your interview?
@cognizant Did someone get this below q for the first tech phone interview?! DM: Product Sense questions, designing a DM followed by 4 SQL questions Can you also give an example for this? - ETL question was to write sql code to load data from staging to target and not a design question really I am curious how did your interview go? Did you get the job?
Following
Cars
Yesterday
1578
Electric cars depreciate 10 times faster than gasoline cars
India
Yesterday
966
What do vegetarian Indians eat for protein?
Tech Industry
Yesterday
677
East Asian Men don’t talk to me bc I’m Vietnamese
Tech Industry
Yesterday
1506
I do tech screens at Google. AMA
India
Yesterday
772
Congress = Muslim league
Pretty much in the same boat, except that the interview hasn’t scheduled yet! We can do a combined prep if you want.
Dm'd you....