Hi! I am Han Lee.
I build and operate machine learning systems. Currently I am building and operating Generative AI applications, search and discovery systems, and LLM models.
In this blog, I journal my thoughts on machine learning engineering, data science, and MLOps.
Occassionally I might write about the technology industry as someone who has sell-side, buy-side, and venture capital investing experiences in the tech sector.
Recent Posts
-
Reasoning Series, Part 1: Understanding GPT-o1
OpenAI's GPT-o1 preview, released in September 2024, introduces "reasoning tokens" to enhance complex problem-solving capabilities. This post explores the model's reasoning process, debunks rumors, and clarifies how users can leverage GPT-o1 for tasks that benefit from deeper thinking.
-
Adapt or Obsolete: The Imperative of Updating Your Priors in the Age of Scaling Compute
In the rapidly evolving fields of AI and NLP, traditional expertise can become obsolete almost overnight. This post explores the shift from hand-crafted features and parsing tasks to end-to-end deep learning models, particularly with the advent of transformers. Learn why it's crucial to continuously update your knowledge and adapt to new technologies, or risk becoming outdated in the face of rapid advancements.
-
Frameworks for LLMs and Compound AI Systems through the Lens of 50 Years of Semiconductor History
The development of compound AI systems, powered by large language models (LLMs) like OpenAI's GPT-o1, marks a transformative shift in AI technology. Drawing on lessons from 50 years of semiconductor history, this post explores how advancements in AI optimization frameworks mirror the evolution from manual gate-level design to automated tools and scalable frameworks. By examining the parallels between hardware design and cognitive computing, we can better understand the future trajectory of LLM systems and the potential for scaling complex AI architectures.