Design Machine Learning Model for Facebook Groups Post Ranking

Quick Overview

This interview question evaluates core ML concepts, assumptions, math intuition, training/evaluation trade-offs, and practical failure modes in a realistic interview setting. A strong answer for Design Machine Learning Model for Facebook Groups Post Ranking states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

Design Machine Learning Model for Facebook Groups Post Ranking

Company: Meta

Role: Data Scientist

Category: Machine Learning

Difficulty: hard

Interview Round: Onsite

##### Scenario Ranking Facebook Groups posts in a user's Newsfeed ##### Question Design a machine-learning model to recommend Facebook Groups posts in Newsfeed. What features, labels, model family, offline/online evaluation, and deployment considerations would you use? ##### Hints Think engagement signals, embeddings, multi-task learning, A/B validation.

Quick Answer: This interview question evaluates core ML concepts, assumptions, math intuition, training/evaluation trade-offs, and practical failure modes in a realistic interview setting. A strong answer for Design Machine Learning Model for Facebook Groups Post Ranking states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

|Home/Machine Learning/Meta
Meta logo
Meta
Aug 4, 2025, 10:55 AM
hardData ScientistOnsiteMachine Learning
4
0

Design Machine Learning Model for Facebook Groups Post Ranking

ML System Design: Ranking Facebook Groups Posts in News Feed

Scenario

You are designing a machine-learning system to recommend posts from Facebook Groups in a user's News Feed. The goal is to maximize user value and healthy engagement while respecting integrity, privacy, and latency constraints.

Assume:

  • Posts can come from groups the user has joined and (optionally) public groups suggested to the user.
  • The recommender will operate within a larger News Feed system and must blend well with other content types.
  • Strict guardrails against spam/low-quality content and negative feedback are required.

Task

Propose a machine-learning approach covering:

  1. Features: What user, group, content, and context features would you use?
  2. Labels/Objectives: What labels and objective(s) would you optimize? How would you handle multiple signals and negative feedback?
  3. Model Family: What modeling architecture(s) would you choose and why?
  4. Offline Evaluation: Which metrics and methodologies would you use to validate offline?
  5. Online Evaluation: How would you run A/B tests and guardrail metrics?
  6. Deployment Considerations: Data/feature pipelines, latency, cold start, integrity, monitoring, exploration, and blending with the broader feed.

Hint: Consider engagement signals, embeddings, multi-task learning, and A/B validation.

Constraints & Assumptions

  • Preserve the scope, facts, inputs, and requested outputs from the prompt above.
  • If the prompt leaves a detail unspecified, state a reasonable assumption before relying on it.
  • Keep the answer interview-ready: concise enough to present, but concrete enough to implement or evaluate.

Clarifying Questions to Ask Guidance

  • Clarify the task, data shape, labels, constraints, and evaluation metric.
  • State assumptions behind the math or modeling technique you choose.
  • Connect theory to practical training, debugging, and deployment implications.

What a Strong Answer Covers Guidance

  • Correct definitions and formulas where the prompt requires them.
  • A practical explanation of how the method behaves on real data.
  • Trade-offs, failure modes, diagnostics, and mitigation strategies.
  • Evaluation choices that match the product or modeling objective.

Follow-up Questions Guidance

  • How would noisy labels, class imbalance, or distribution shift affect the answer?
  • What would you monitor after deployment?
  • Which baseline would you compare against first?
Loading comments...