English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Scaling the Instagram Explore Recommendations System (Meta Engineering, August 2023)

Forum topic · 小凯 · 2026-07-05

Summary

This Meta Engineering blog post (August 9, 2023) describes how Meta scaled the Instagram Explore recommendations system to handle billions of daily user interactions. It covers the retrieval-and-ranking pipeline behind Explore, which surfaces relevant media from an enormous candidate pool, and the architectural changes needed to grow embedding-based retrieval, candidate generation, and lightweight ranking models at scale. Key themes include moving to more efficient multi-stage ranking, improving the use of embeddings and ANN indexes, batching and hardware optimizations, and reducing the cost and latency of serving personalized content to hundreds of millions of users. The post explains how Meta balanced model complexity against serving efficiency, adopting techniques such as model compression and better feature engineering to deliver higher-quality recommendations without proportional infrastructure growth. Engineers and researchers working on large-scale recommender systems will find practical insights on pipeline design, scalability trade-offs, and production metrics for one of the world's largest recommendation workloads. Original source: https://engineering.fb.com/2023/08/09/ml-applications/scaling-instagram-explore-recommendations-system/

Scaling the Instagram Explore Recommendations System (Meta Engineering, August 2023)

  • Source: Meta Engineering Blog, Aug 9, 2023
  • Type: Industry engineering blog
  • Key points

  • Instagram Explore is Meta's recommendation surface for discovering new content outside a user's existing network, serving hundreds of millions of users and requiring ranking over billions of candidate media items.
  • The system uses a classic multi-stage pipeline: candidate generation (embedding-based retrieval over a huge corpus) followed by multi-stage ranking that progressively applies more complex models to fewer candidates.
  • Meta describes scaling efforts focused on making each stage more efficient so that heavier models and richer features can be introduced without linear infrastructure cost growth.
  • Engineering optimizations include improved embedding retrieval infrastructure, batched and distributed serving, and careful trade-offs between model size, feature count, and p99 latency budgets.
  • The post emphasizes that quality improvements must be validated with online experiments (A/B tests) measuring engagement and satisfaction, not just offline ranking metrics.
  • Architecture and pipeline

    1. Candidate generation: media candidates are retrieved using embeddings and approximate nearest neighbor search, balancing recall against computational cost. 2. Ranking stages: candidates pass through lightweight rankers first; only the top subset reaches heavier models, keeping total serving cost manageable. 3. Feature infrastructure: user, item, and contextual features feed the ranking models; scaling required more efficient feature computation and storage. 4. Serving efficiency: batching, model compression, and hardware-aware design reduce per-query latency and compute, enabling scale to billions of daily requests.

    Takeaways for practitioners

  • Efficiency work (retrieval infrastructure, staged ranking) is what buys headroom for model quality improvements at scale.
  • Latency and cost budgets are hard constraints: heavier models are only viable where the staged architecture limits their invocation.
  • Offline metrics must be cross-checked against online experiments to ensure perceived recommendation quality actually improves.
> Note: This page is a structured English rendering of a Chinese forum post that summarized the original Meta blog. Quantitative figures should be verified against the original article.

Tags

#recommendation-systems#instagram#meta#ranking#embedding-retrieval#machine-learning#scalability

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178208499