Eugene Yan
2025年3月18日 · Eugene Yan is a senior applied scientist at Amazon building machine learning systems that serve customers at scale. Previously, he led ML/AI at Alibaba, Lazada, and a Healthtech Series A. He writes and speaks about RecSys, LLMs, and engineering at eugeneyan.com.
Patterns for Building LLM-based Systems & Products - Eugene Yan
“Obsidian-Copilot: An Assistant for Writing & Reflecting.” eugeneyan.com, (2023). Bojanowski, Piotr, et al. “Enriching word vectors with subword information.”
About • Eugene Yan
Eugene Yan is a Senior Applied Scientist at Amazon building machine learning systems that serve customers at scale. Previously, he led ML/AI at Alibaba, Lazada, and a Healthtech …
Start Here • Eugene Yan
Key themes I write about including machine learning systems & techniques, ML in production, data science methodology, writing, learning, and career.
Some Intuition on Attention and the Transformer - Eugene Yan
What's the big deal, intuition on query-key-value vectors, multiple heads, multiple layers, and more.
Real-time Machine Learning For Recommendations - Eugene Yan
How does real-time ML look in practice? We discuss real-time recommenders, drawing on my experience & industry papers: • When does it make sense? When does it not? • How have China & US companies implemented them? • How do we design & build an MVP? https://t.co/hSfRr3q2ie — Eugene Yan (@eugeneyan) January 12, 2021
What I Wish I Knew About Onboarding Effectively - Eugene Yan
I hope this advice makes your first 100 days more effective. All the best with your new role! Thanks to Yang Xinyi for reading drafts of this. If you found this useful, please cite this write-up …
Georgia Tech's OMSCS FAQ (based on my experience) - Eugene Yan
I’ve received many qns about Georgia Tech’s OMSCS, especially since graduation. • Why further studies? Why OMSCS? • How can I get accepted? How much time needed? • What classes …
Evaluating the Effectiveness of LLM-Evaluators (aka LLM-as-Judge)
LLM-evaluators, also known as “LLM-as-a-Judge”, are large language models (LLMs) that evaluate the quality of another LLM’s response to an instruction or query. Their growing adoption is partly driven by necessity. LLMs can now solve increasingly complex and open-ended tasks such as long-form summarization, translation, and multi-turn dialogue. As a result, …
How to Match LLM Patterns to Problems - Eugene Yan
After my previous write-up on LLM patterns, I’ve received questions on how to match those patterns to various LLM problems. Thus, in this follow-up, we’ll discuss some potential problems faced when using LLMs and the patterns that help mitigate them. External vs. internal LLMs, data vs. non-data patterns Before we dive into it, I think it’s helpful to distinguish between external …