Inside LinkedIn’s Generative AI Cookbook: How It Scaled People Search to 1.3 Billion Users

Why It Matters

The rollout shows that enterprise‑scale generative AI can be delivered to over a billion users through pragmatic, repeatable pipelines rather than hype‑driven agents, offering a template for other large tech firms. It also sharpens LinkedIn’s core value proposition by surfacing more relevant professional connections, paving the way for future AI‑driven networking tools.

Summary

LinkedIn has rolled out an AI‑powered people search that lets users pose natural‑language queries—e.g., “Who is knowledgeable about curing cancer?”—to a graph of 1.3 billion members. The system combines a 7‑billion‑parameter policy model distilled into a 1.7 B teacher and a 220 M student model, with an 8 B retrieval model running on GPU‑based infrastructure and a reinforcement‑learning summarizer that cuts input size 20‑fold, delivering a ten‑fold boost in ranking throughput. The launch builds on a “cookbook” of staged distillation, synthetic data generation and aggressive pruning first proven in LinkedIn’s AI job‑search product, which raised hiring odds for non‑degree seekers by 10 %.