# Connected Content Retriever: Dense Graph Edge Features Powering Pre-Ranking at LinkedIn

arXiv: https://arxiv.org/abs/2609.22441 · HTML: https://mlsanity.com/paper/2609.22441/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 18 references that are in the map.
- First posted on arXiv 18 Sep 2026.

## Builds on (most cited first, 9 of 9)

- [Beyond Two-Tower Matching: Learning Sparse Retrievable Cross-Interactions for Recommendation](https://mlsanity.com/paper/2311.18213/) (2023)
- [SilverTorch: A Unified Model-based System to Democratize Large-Scale Recommendation on GPUs](https://mlsanity.com/paper/2511.14881/) (2025)
- [Large Scale Retrieval for the LinkedIn Feed using Causal Language Models](https://mlsanity.com/paper/2510.14223/) (2025)
- [LinkedIn Post Embeddings: Industrial Scale Embedding Generation and Usage across LinkedIn](https://mlsanity.com/paper/2405.11344/) (2024)
- [InteractRank: Personalized Web-Scale Search Pre-Ranking with Cross Interaction Features](https://mlsanity.com/paper/2504.06609/) (2025)
- [HIT Model: A Hierarchical Interaction-Enhanced Two-Tower Model for Pre-Ranking Systems](https://mlsanity.com/paper/2505.19849/) (2025)
- [Equip Pre-ranking with Target Attention by Residual Quantization](https://mlsanity.com/paper/2509.16931/) (2025)
- [AIF: Asynchronous Inference Framework for Cost-Effective Pre-Ranking](https://mlsanity.com/paper/2511.12934/) (2025)
- [An Industrial-Scale Sequential Recommender for LinkedIn Feed Ranking](https://mlsanity.com/paper/2602.12354/) (2026)

## Abstract (arXiv)

In large-scale recommendation systems like the LinkedIn Feed, content generated by a member's network (connections and follows) makes up over 70% of impressions and engagement. It is therefore essential that the pre-ranking layer forwards the best possible few hundred candidates to the ranking layer. LinkedIn's professional knowledge graph carries engagement signals across both the first degree network (connections and follows) and the second-degree network: posts that a 1st-degree connection reacted to, commented on or reshared but did not author (a.k.a. stranger viral). Due to this fan out, the resulting candidate index exceeds one billion; selection of activities from the viewer's network narrows it down to roughly tens of thousands of activities that must be scored within a 120 ms p99 latency budget. We present Connected Content Retriever (CC Retriever), a pre-ranking system that scores these candidates with a full deep ranking model on GPUs at low latency. At its core is a sorted-search GPU primitive that joins dense graph affinity features (viewer to author) with document level features stored on the GPU at runtime in 5-10 ms. The shift to GPU served scoring enabled a 50x scale up of the ranking model's parameters and delivered a +2.5% lift in content time spent on the LinkedIn Feed in online experiments, significantly higher than the typical gains observed in LinkedIn Feed experiments. In this work, we describe the feature set we leverage from LinkedIn's economic graph and the model architecture used for scoring, with a particular emphasis on the online system that scales the stack.