Practice real data science interviews tuned by the interviewers who actually run them.

Facebook's News Feed team shipped a ranking algorithm change last week. Since then, comments per user are down 12% but time spent is up 5%. The VP wants to know: is this a problem?

Facebook is considering removing visible like counts from posts. Instagram already tested hiding likes in several markets. You need to think through how to measure the full impact of this change across users, creators, and the platform ecosystem.

The Facebook Newsfeed team is trying to understand whether users see enough friend content in their feed. The balance between friend posts, public content, and recommended content is critical to Facebook's value proposition. You need to figure out how to measure and optimize this.

The Reels team has flagged an alarming trend: 'average reels watched per session' has dropped precipitously over the past week. The VP wants answers by end of day. You need to diagnose what's happening and whether it's actually a problem.

The PM responsible for Facebook Events has a new idea to drive engagement. When your friends mark that they'll attend an event, you'll receive a notification. How would you measure the success of this notification?

Facebook is creating a notification that alerts users when their Marketplace listing is about to expire. Your PM wants to measure the quality and effectiveness of this notification before a broader rollout.

Facebook is launching a new product called Facebook Restaurants — a Marketplace-like experience for discovering, reviewing, and ordering from local restaurants. You need to define the target users, success metrics, and measurement plan.

Facebook is launching a group chat feature similar to Discord channels within Facebook Groups. The goal is to increase real-time engagement within communities. You need to define success and guardrail metrics for the launch.

Meta's leadership has been pushing teams to optimize for user wellbeing, not just engagement. You've been asked to develop a framework for measuring whether a feature actually makes users happier — a much harder question than whether it drives engagement.

Instagram's Reels product has been live for over a year. Leadership wants a comprehensive health check: is Reels actually succeeding, or is it just cannibalizing other Instagram surfaces? You need to define the metric framework for evaluating Reels as a product line.

The Instagram Monetization team wants to double the amount of ads shown on Instagram — it's the quickest way to nearly double revenue overnight. How would you determine the optimal ad-load for Instagram?

Facebook Search has observed a 10% increase in clicks on Events. Before celebrating, you need to figure out whether this is a good sign, a bad sign, or something more nuanced. The PM wants your analysis.

Facebook Groups wants to test a new feature: AI-generated discussion prompts posted automatically in groups with declining activity. The goal is to re-engage dormant groups. How would you design this experiment?

You're a data scientist at a food delivery company. The operations team ran a test where they gave Dashers a guaranteed minimum per delivery in a specific market. Results are in — but they're mixed.

Facebook is integrating Reels into the main Newsfeed experience. This is a major surface change that could boost short-form video consumption but risks cannibalizing existing content types. You need to design the experiment and measurement plan.

The Messenger team wants to improve the video call experience. They have several ideas — better connection quality, AR filters, screen sharing, and background blur. You need to design an experiment to test improvements and define what success looks like.

Instagram's leadership is debating where to invest their next engineering cycle: doubling down on Reels (short-form video) or revamping Stories (ephemeral content). You're asked to present a data-driven framework for making this decision.

Facebook is considering adding a 'restaurants you may like' recommender in the News Feed. This would be a new content type competing for feed real estate alongside posts, Reels, ads, and stories. You need to design the system and measurement plan.

WhatsApp is seeing a surge in spam messages across both individual and group chats. You've been asked to design a spam detection system. The catch: WhatsApp is end-to-end encrypted, which limits what data you can use.

Meta's integrity team needs to estimate the prevalence of misinformation in Facebook Stories. Unlike the News Feed, Stories are ephemeral (24-hour lifespan) and often contain images, video, and text overlays — making automated detection harder. You need to design the measurement approach.