Meta Interview Questions
& Process

I applied online and got a recruiter call after about two weeks to chat about my background and if I'd be a good fit. Then I had a technical phone interview where I had to code some problems, like LeetCode medium-hard level, focusing on arrays and graphs. After that, I did a virtual onsite interview that had 4 parts: two coding rounds, one ML system design round where they asked about recommendation or ranking systems, and one behavioral round. The interviewers were really professional and let me ask questions at the end. They told me the results about a week after the onsite.

The interview process consisted of 5 interviews: 1) Hiring Manager with some system design questions 2) Program Sense: Assessing the ability to manage complex projects from end-to-end, including early-stage roadmap creation, problem definitions, scope and requirements gathering, project planning, problem-solving, and risk mitigation. 3) Results Agility: Evaluating the results in new or unstructured situations. Responding to tough challenges. 4) TPM System Design: Evaluating understanding of the product by clearly identifying requirements and use cases (Product Sense) and going into the technical details (e.g. design, scaling, product enhancement, e2e system design, tradeoffs/benefits, solutions on how to scale, define MVP). 5) Product Sense: Evaluating the understanding of basic KPIs and metrics, and what it takes to launch a product; define business/performance related metrics.

I recently interviewed for a Software Engineer role. The technical interviewers were professional, but the recruiting process was one of the most inefficient and deceptive experiences I’ve had in the industry. 1. Misleading and "Ghost-Rejection" Emailing Style: Two days after the full loop interview, the recruiter sent an email titled "Thank You for Your Time Today!" packed with overly enthusiastic language ("huge thank you," "enthusiasm," "here for you every step of the way"). For any experienced candidate, this reads like a "soft" rejection template, yet it was used just to tell me to wait for their result. It came across as a deliberate attempt to manipulate the candidate’s expectations instead of giving a clear, professional status update. 2. Pointless Sync Call for a 30-Second Rejection: Even though they had the result ready by Friday, the recruiter insisted on setting up a formal "sync call" for the following Tuesday morning. I had to rearrange my own work schedule and reserve a private room at my current office, only for the recruiter to spend 30 seconds reading a generic rejection script. When I shared feedback that this could—and should—have been handled over email to respect everyone’s time, the recruiter doubled down, saying they wanted to "show gratitude." There is no gratitude in wasting a candidate’s productive working hours for an update that fits into two sentences. 3. The Outcome: After all this "high-touch" (really: high-friction) communication, it ended in a standard 12-month cool-off period. Advice to Management Efficiency is a core value at Meta, but it is clearly missing in the recruiting department. Stop forcing "sync calls" for rejections. It is not "personal"; it is disruptive. If you have a decision on Friday, send the email on Friday. Respect candidates’ working hours and their professional time.

Initial Screening 3 SQL 3 Python SQL: Star schema, sales transaction fact, book author etc dims * Identify authors who have published at least 5 books * Calculate the percentage of total sales completed on the same day the customer registered * Find customers who purchased 3 or more books on both the first and last day of sales, excluding those with only one transaction * Find the IDs of the top 5 customers, ordered by average payment per book made by the people they invited * Find the total number of authors registered with the bookstore. What percentage of them have a website URL that contains ".com", and what percentage never made a sale? * What was the total value of sales and the number of unique paying customers, grouped and sorted in descending order by payment type? * Having a transaction table find the sum of total orders and the count of unique customers Python: * Calculate the average book price from a list of prices * The bookstore gathered a list of customer comments from each shop location and wants to find the most common comment across all locations (ignoring duplicates from the same location). If multiple comments appear the same number of times, return any one of them. * Search an element in an unsorted list * Given a list of workshops, return the largest number of classes that were hosted in total across any consecutive years that had at least one workshop each.