← Meta Interview Insights

Meta·Data Scientist·Technical Phone Screen·Senior

Senior
May 2026

Summary

Meta DS interview centered on a single meaty ML system design question about restaurant recommendations in a social feed. No behavioral rounds mentioned, just the one technical deep dive. Felt like they wanted to see how you'd think end-to-end rather than just spit out model names.

Questions Asked (1)

Q1

Design a recommendation system that surfaces relevant local restaurant pages to users in their news feed (organic, not ads). Walk through your data sources, feature choices, model architecture, real-time ranking, and how you'd evaluate the system offline and online.

System DesignA/B Testing & ExperimentationProduct Analytics & Metrics
Author's notes

This is a big question and I underestimated how far they'd want me to go.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the product goal and constraints (e.g., organic feed, local restaurant discovery, user engagement). Then walk through the end-to-end system: data sources, feature engineering, model architecture, real-time ranking, and evaluation. Emphasize how you balance relevance, freshness, and diversity while addressing cold-start and scalability.

Pro tip: Tie every design choice back to a measurable objective (e.g., CTR, dwell time, or restaurant visits) and explicitly discuss trade-offs like latency vs. model complexity. Show awareness of Meta's scale by mentioning distributed training and serving, and how you'd handle billions of daily impressions.

1. Clarify Goals and Constraints

Define the problem: what does 'relevant' mean? Identify key metrics (e.g., clicks, saves, direction requests) and constraints (latency, scale, privacy). Confirm the feed context and user intent.

2. Data Sources and Feature Engineering

List data: user profiles, historical interactions, restaurant attributes (cuisine, location, ratings), contextual signals (time, location, device). Engineer features: user-restaurant affinity, distance, popularity, freshness, and embeddings.

3. Model Architecture

Propose a two-stage architecture: candidate generation (e.g., matrix factorization, two-tower neural network) to retrieve hundreds of candidates, followed by a ranking model (e.g., deep neural network with wide & deep or DLRM) to score and order them.

4. Real-Time Ranking and Serving

Describe how to serve predictions with low latency: precompute embeddings, use approximate nearest neighbor search for retrieval, and a lightweight ranking model. Incorporate real-time features (e.g., user's current location) via feature store.

5. Evaluation: Offline and Online

Offline: use historical data with metrics like NDCG, recall@k, and AUC; conduct counterfactual or off-policy evaluation. Online: A/B test with guardrail metrics (CTR, dwell time, diversity) and long-term holdouts to measure retention.

Key Points to Mention

  • Two-stage retrieval and ranking architecture to handle scale
  • Feature engineering: user-restaurant affinity, contextual features (location, time), and restaurant embeddings
  • Handling cold-start for new users and restaurants via content-based features and exploration
  • Real-time serving with low latency using approximate nearest neighbor search and feature stores
  • Offline evaluation metrics (NDCG, recall@k) and online A/B testing with guardrail metrics
  • Diversity and freshness in ranking to avoid filter bubbles and maintain user engagement

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.