在 Netflix 扩展 LLM 后训练

[

Netflix TechBlog

](https://netflixtechblog.com/?source=post_page---publication_nav-2615bd06b42e-0046f8790194---------------------------------------)

[

Netflix TechBlog

](https://netflixtechblog.com/?source=post_page---post_publication_sidebar-2615bd06b42e-0046f8790194---------------------------------------)

Learn about Netflix’s world class engineering efforts, company culture, product developments and more.

了解 Netflix 的世界一流工程努力、公司文化、产品开发等。

Baolin Li, Lingyi Liu, Binh Tang, Shaojing Li

Introduction

引言

Pre-training gives Large Language Models (LLMs) broad linguistic ability and general world knowledge, but post-training is the phase that actually aligns them to concrete intents, domain constraints, and the reliability requirements of production environments. At Netflix, we are exploring how LLMs can enable new member experiences across recommendation, personalization, and search, which requires adapting generic foundation models so they can better reflect our catalog and the nuances of member interaction histories. At Netflix scale, ==post-training quickly becomes an engineering problem as much as a modeling one==: building and operating complex data pipelines, coordinating distributed state across multi-node GPU clusters, and orchestrating workflows that interleave training and inference. This blog describes the architecture and engineering philosophy of our internal Post-Training Framework, built by the AI Platform team to hide infrastructure complexity so researchers and model developers can focus on model innovation — not ==distributed== systems plumbing.

预训练赋予 Large Language Models (LLMs) 广泛的语言能力和一般世界知识，但 post-training 是真正将它们对齐到具体意图、领域约束以及生产环境可靠性要求的阶段。在 Netflix，我们正在探索 LLMs 如何在推荐、个性化...