Interview Questions

Ranj A. - "In the Transformer architecture, the decoder differs from the encoder primarily in its additional mechanisms designed to handle autoregressive sequence generation. Here's a breakdown of the key differences: Self-Attention Mechanism: Encoder: The encoder has a standard self-attention mechanism that allows each token to attend to all other tokens in the input sequence. Decoder: The decoder has two types of self-attention. The first is the same as in the encoder, but the second is mas"See full answer

Esha - "Clarifying Questions What do we mean by ‘improve’? Is there a certain goal that we are trying to attain? Is there a target user that we are going for? Any constraints or things to keep in mind when building this? Assumptions Only focused on google docs and not collaboration on other G-suite products such as PPT or sheets What is Google Docs? Web-based productivity tool that allows you to access your files from any computer through your gmail account. You can also share and"See full answer

Aruna D. - "Assumption #1: the deal is signed and the customer now will being the process of moving to GCP Assumption #2 : the deal includes Iaas as well as Paas (App Engine) - moving apps from on prem to cloud Objective : Design a product that will help the customer onboard and successfully plan the move to GCP. Customer is using on premise and is new to cloud. Lets assume the tool should include the roadmap planning too ie decisions such as moving all vs some applications etc . As in any onboarding to"See full answer

Erjan G. - "1) select avg(session) from table where session> 180 2) select round(sessiontime/300)*300 as sessionbin, count() as sessioncount from table group by round(sessiontime/300)300 order by session_bin 3) SELECT t1.country AS country_a, t2.country AS country_b FROM ( SELECT country, COUNT(*) AS session_count FROM yourtablename GROUP BY country ) AS t1 JOIN ( SELECT country, COUNT(*) AS session_count FROM yourtablename `GROUP BY countr"See full answer

Meagan R. - "Collaborate with the team to determine which team's product will move the needle the most in achieving the strategic goals and prioritize that work."See full answer

Hari S. - "Clarifying questions: Is this a standalone product or a Meta integrated product? Which region is this for? US or another region or global? Is this for a county run library or a more informal group community library? Assumptions: It's a meta integrated product US based system group run community library Let's see if a lending library fits into Meta's company mission to build community and bring people closer together. Yes, a lending library does build community by helping peop"See full answer

Valentin B. - "Let me first ask some clarification questions / make assumptions: Any specific geographic region? => USA How do we measure bandwith? a. Bandwith is the maximum internet speed required for our google maps servers to support all google maps traffic in the USA. At which point do we measure bandwith? => at the “entrance” of each google server center Do we talk about maximum bandwith required? a. Yes, the maximum BB required to support google maps in the USA (on side of Google) Outline: O"See full answer

Anonymous Puma - "Assuming for the US Segment by size of businessNumber of employees per business sizePrice per employee=Revenue Price per employee should decrease with a larger employee base. 30M businesses in the US Small Business 60% of businesses are small=18M Average number of employees=50 Price per employee=$5 Small Business Revenue=$2.7B Medium Business 30% of them are medium=9M Average number of employees=1000 Price per employee=$3 Medium Business Revenue=$8.1B Large Business 10% of th"See full answer

Michael F. - "As an interviewer, I have asked this question to candidates in the past. Here are the major topics I am looking for in an interview The candidate should understand that there are ways of measuring the loss of a particular clustering. For example, we can take the average distance of each point to it's cluster center. The candidate should understand that this loss will always decrease as the number of clusters increases. For that reason, we can't just pick the value of K that minimizes the l"See full answer

Dan D. - "It would be nice if the final Google Docs link can be shared somewhere, so we can access it for reference with ease."See full answer

R K. - "Clarifying questions: Can we pick any business product across CapOne spectrum? any preference? - Assume pick any Are there any specific geographies you'd like to focus on? Are there any specific customer segments you'd like to focus on? Are you able to provide any information on the market/competitive analysis? Or I can just make reasonable assumptions along the way. Confirming we care about revenue? not profit per customer? Lay out the framework: I'd like to focus on the followi"See full answer

Dinar M. - "Even more faster and vectorized version, using np.linalg.norm - to avoid loop and np.argpartition to select lowest k. We dont need to sort whole array - we need to be sure that first k elements are lower than the rest. import numpy as np def knn(Xtrain, ytrain, X_new, k): distances = np.linalg.norm(Xtrain - Xnew, axis=1) k_indices = np.argpartition(distances, k)[:k] # O(N) selection instead of O(N log N) sort return int(np.sum(ytrain[kindices]) > k / 2.0) `"See full answer

Manas M. - "I'd ask Clarification questions here before proceeding with designing a CDN. The questions would be. What is the Use Case with multiple questions in it What sort of Content do you want to serve - Static files or Streaming Media or trending content? How many regions do you want to cover Do you want to prioritize speed and the Latest file availability or do you want to prioritize time-based caching per region? Budget constraints(push cdn's can be cheaper). Whether"See full answer