The Curious Engineer

All blogs / AI / ML

Han, Not Solo

leehanchung.github.io

Essays by Han Lee on machine learning systems, AI evaluation, agent infrastructure, search, recommendation systems, and engineering practice.

AI / MLSoftware / engineering practiceEngineering management / careers

’19’20’21’22’23’24’25’26
Posts
78 from 26 Jun 2019 to 20 Sep 2026
Posting
monthly (12 in the last year)
Kind
personal blog
Feed
https://leehanchung.github.io/feed.xml

2026 11

  1. Jev and the return of AI/ML Engineering
  2. Stop Being Permanent Underclass. Implement a Token Factory.
  3. Hidden Technical Debt of AI Systems: Agent Evaluation Infrastructure
  4. Data Aggregation Is Not a Moat
  5. Hidden Technical Debt of AI Systems: Agent Harness
  6. Don't Outsource Your Understanding
  7. Hidden Technical Debt of AI Systems: Agent Runtime
  8. "Determinism" is the Biggest Cope in AI Adoption
  9. The AI Great Leap Forward
  10. A Taxonomy of RL Environments for LLM Agents
  11. It's-a Me, Agentic AI

2025 20

  1. Databricks' Strategic Playbook: Reynold Xin on Growth, AI, and the Future of Data Infrastructure
  2. Claude Agent Skills: A First Principles Deep Dive
  3. Building for IT 3.0: Why Agent-Native Products Will Crush SaaS Click-Ops
  4. Enterprise AI Transformation: The 4-Set Framework for IT 3.0
  5. Statistics for AI/ML, Part 4: pass@k and Unbiased Estimator
  6. How AI Tools Are Reshaping Software Development Team Responsibilities
  7. Software Engineering for Data Scientists, Part 1: Pydantic Is All You Need for Poor Performance Spaghetti Code
  8. No Code, Low Code, Real Code
  9. MCP is not REST API
  10. Prompt Deployment Goes Wrong: xAI Grok's obsession with White Genocide
  11. Agents Are Workflows
  12. Vibe Coding 101 for Software Engineers
  13. When Prompt Deployment Goes Wrong: MLOps Lessons from ChatGPT’s 'Sycophantic' Rollback
  14. Pro Tips: Model Context Protocol developement on Windows
  15. Walkthrough: Setting Up Claude Code on Amazon Bedrock
  16. Poking Around Claude Code
  17. Solve AI Engineering Slop with Machine Learning
  18. Statistics for AI/ML, Part 3: Measuring Classification Agreements Between Agents
  19. Statistics for AI/ML, Part 2: Measuring Ranked Agreements Between Agents
  20. The Differences between Deep Research, Deep Research, and Deep Research

2024 19

  1. Statistics for AI/ML, Part 1: Adding Confidence Interval to Your Aggregation Statistics
  2. Reasoning Series, Part 4: Reasoning with Compound AI Systems and Post-Training
  3. Reasoning Series, Part 3: Reasoning with Prompt Engineering and Sampling
  4. What The Hype and Reality of Agents: Resurgence of the Past, with LLM Flavors
  5. Reasoning Series, Part 2: Reasoning and Inference Time Scaling
  6. Reasoning Series, Part 1: Understanding OpenAI o1
  7. Updating Your Priors as Compute Scales
  8. Frameworks for LLMs and Compound AI Systems through the Lens of 50 Years of Semiconductor History
  9. AI Winter Is Coming
  10. Seven Key Metrics Every SaaS Business Needs to Thrive in the AI Era
  11. LLM-as-a-Judge: Rethinking Model-Based Evaluations in Text Generation
  12. AI System UX: From Chatbots to Agents
  13. Stop Following Tech Influencers
  14. First Principles in Evaluating LLM Systems
  15. NSFW: GPT-4o Tokenizer
  16. Tools for Large Language Model Agents
  17. Information Trifecta
  18. Service as Software
  19. Subject: search lingo

2023 4

  1. Decoding LLM Pricing: The Cost of Input and Output Sequences in Generative AI
  2. Data Leakage in WizardCoder: A LeetCode Reality Check
  3. CUDA x WSL2
  4. Prompt Engineering Guide

2022 2

  1. Notes from SIGIR 2022
  2. Notes from ACM RecSys 2022

2021 14

  1. Notes on from EMNLP 2021
  2. AWS Lambda: Keeping Functions Warm
  3. AWS Sagemaker: Persist Custom Conda Environment
  4. List All Columns in a Table in Redshift
  5. Pandas fillna for Categorical Columns
  6. Syncing Conda Environments with requirements.txt
  7. XGBoost Hyperparameters Tuning
  8. SageMaker: Automatically Stop Notebook and Studio Instances When Idle
  9. Getting Months from a Date Range in Pandas
  10. Tensorflow and Friends
  11. Boruta Algorithm
  12. Weight of Evidence and Information Value
  13. Managing Multiple AWS and Git profiles
  14. Feature Selection and Dimensionality Reduction

2020 3

  1. Evaluation of Clustering Algorithms for Information Retrieval
  2. How Tik-Tok Wins The Social Media Recommendation System War
  3. AWS CPU Instance Numpy Benchmarking Roundup

2019 5

  1. Serve Keras Model with Tensorflow Serving On AWS Sagemaker
  2. Serve Keras Model with Tensorflow Serving
  3. Pushing Docker Images to AWS Elastic Container Registery (ECR)
  4. Tuning XGBoost Parameters
  5. GDP Growth And Stock Market Return Correlates Less Than You Think