Interview Questions

Jay K. - "First, let’s discuss why this is important for any platform business Think of platform businesses like Amazon or YouTube as matchmakers. They connect people who want something with those who provide it, whether it's products or content. But here's the catch: users trust these platforms to make sure they're getting good stuff. They want genuine products, safe content, and no scams. So, platforms need to be like guardians. They check suppliers and content creators to make sure they're not"See full answer

Umang S. - "Context - I joined a large public-facing service as a PM midway through its development. Situation - Due to the SOPs of the company the team already had a metrics framework. That included your standard DAUs, Retention and Acquisition Metrics Concern - As SOP metrics were publicly accepted, the team did not internalise what success actually means to the product they are developing, as the actual value was not being encapsulated in the framework This was evident in the show and tells as the"See full answer

Apurv K. - "!! FEEDBACK WELCOME !! Clarify question By art we mean visual art, e.g. paintings? - Yes Is this a solution to create art on Instagram or to showcase existing art on Instagram? - Create art Why build this, Engagement or Acquisition? - Engagement Reframe: We want to increase Instagram engagement through creation and sharing of art with followers. Talking about the space The Internet has over time continuously reduced barriers to creative expression. From blogging to Twitter"See full answer

Usman B. - "System Components Data Collection Layer Posts with hashtags are asynchronously sent to Kafka topics Each message contains: hashtag, timestamp, userid, postid Multiple Kafka partitions ensure scalability and fault tolerance Processing Layer Apache Flink processes streams in real-time Implements sliding window aggregation (1hr, 24hr, 7d windows) Calculates topic popularity using weighted metrics: Post count User engagement (likes, comments) Unique user"See full answer

Ash J. - "Clarify: I will assume that misinformation includes posts, videos and pictures and that sources could be Groups/ Pages and Individuals. High level structure: Mission, Information funnel and drop-off points, Metrics & Trade-offs, Mitigations Mission/ Strategy: Facebook's mission is to bring the world closer together and empower people to build community. Reliable information is key to fostering a healthy community. In light of recent world events and public sentiment, it is important f"See full answer

Kartikeya N. - "Clarifying questions: What city/ country are we talking about? Indian metro. What all counts as a grocery store? Do have to cover dark stores as well which are not open to public? Only consider all grocery stores where public pick up is allowed. Do we have to include hyper markets/ super markets as well or only smaller stores? Include both. Okay, so I would be considering the Indian metro of Bangalore here. The overall equation is , **total number of grocery stores = Number of small grocer"See full answer

Hans - "Described a scenario where I joined an established project as an EM Typical medium sized team - 6 devs, 2 testers, PO, PM (part time) Scrum master (very hands off) The product was well established and already being used by circa 30000 single country based users for the companies own products (personal banking capabilities) We were due to go live with a BIG change about 2 months after I joined - this change meant a user could manage multiple personal banking accounts from the 1 interface. We w"See full answer

Anonymous Roadrunner - "-- Write your query here select id, (case when p_id is null then 'Root' when pid in (select id from treenode_table) and id in (select pid from treenode_table) then 'Inner' else 'Leaf' end) as node_types from treenodetable order by 1; `"See full answer

Robert H. - "The addition of an intermediate "sanitization" ML is something Neeraj used in the Uber Eats design and again seems kind of outside the scope here and redundant. This can simply be built into the AI response model to save a step. It's not clear what benefit this step provides, and he basically said we should have it "just because it would be good" and there's no concrete reasoning why to include it. Adding a Kafka queue to handle the thumbs-down ratings? For what purpose do you need a queue othe"See full answer

Hashmina H. - "Clarifying Questions: Do we want to target a particular geography or location or should it be done globally? Assume global Is there any particular platform and device we are targeting (App/web) or should it be cross platforms? Cross platforms Can I think about it as Product led solutions or do you want to consider marketing alternatives as well? Let's start with Product led solutions & later if time permits, we can touch on the marketing side as well Are there any constra"See full answer

Philip C. - "Initial Assumptions I’m going to assume that this is a blue sky question, and there aren’t any constraints in terms of geo, funding, etc. With that in mind, one deeply unoptimized product is… a parking meter. I’ll assume for the sake of this answer that we’re focusing on parking in an urban environment since that’s where most parking-related friction exists. I’m also going to assume that we’re a startup and don’t have any special strengths/ weaknesses that we need to account for"See full answer

Karthik R. - "Good Discussion on the distributed messaging queues (Complex topic with lot of nuances) Liked the mind-map style drawing of requirements and metrics capture Touched on different types of queue styles (point to point, pub-sub, fan-out/fan-in) Storage and WAL usage was interesting Some distributed queue challenges that could be helpful to highlight / expand are: Message guarantee / semantics - Ordering of messages across different servers Replication Master slave architecture or Pe"See full answer

Mike Z. - "Clarifying questions Scope: Just Reels in Facebook, or Instagram too? Mobile and desktop? Goal: Any specific goal we know of or up to me to decide? Framework/Overview Strategy Metric options with pros/cons North Star Metric (NSM) Risks & countermetrics Strategy / The Why: we want to give people … Meta: the power to build community + Bring people closer together Reels: new form of self-expression / entertainment with short form video and pro editing tools Ads in Re"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

Steven M. - "Clarifying questions: I assume we're focused on the consumer side of this? ** Yes I am assuming post booking begins immediately after booking a ticket and ends when the user leaves the airport at their destination. Does that sound right? ** Yes Goals: United Airlines wants to get its passengers to their destination seamlessly, safely, and on time. Users: I think about user groups in 3 main buckets. 1) 24 hours following booking 2) >24 following booking through 24 hours before boarding "See full answer

Will P. - "There are several categories of ads that users may encounter on YouTube content pages, depending on the user's device and browser. To narrow the scope of my answer, I am going to assume this is a non-skippable, midroll video ad that is 15 seconds long, and shown to a non-premium desktop user in the US. The next step is to clarify the motivations behind the proposed feature. Are non-relevant ads harmful to the overall user satisfaction? Do users engage more with relevant ads? Are advertisers mak"See full answer

S R. - "Wrote up a simple cache system using python dict. Added TTL requirement. Then went into code-level concurrency issues for the cache."See full answer

David I. - "WITH filtered_posts AS ( SELECT p.user_id, p.issuccessfulpost FROM post p WHERE p.postdate >= '2023-11-01' AND p.postdate < '2023-12-01' ), post_summary AS ( SELECT pu.user_type, COUNT(*) AS post_attempt, SUM(CASE WHEN fp.issuccessfulpost = 1 THEN 1 ELSE 0 END) AS post_success FROM filtered_posts fp JOIN postuser pu ON fp.userid = pu.user_id GROUP BY pu.user_type ) SELECT user_type, post_success, post_attempt, CAST(postsuccess AS FLOAT) / postattempt AS postsuccessrate FROM po"See full answer