AI
makeMoE: Implement a Sparse Mixture of Experts Language Model from Scratch
nanoMoE: Mixture-of-Experts (MoE) LLMs from Scratch in PyTorch
LLM-Powered Time-Series Analysis
Understanding Convolutional Neural Networks (CNNs) Through Excel 通过 Excel 理解卷积神经网络
How to Build an Over-Engineered Retrieval System 如何构建一个过度设计的检索系统
How agents can use filesystems for context engineering
用 AI“打开”金融市场黑盒：微软亚洲研究院如何构建订单级仿真引擎
20x Faster TRL Fine-tuning with RapidFire AI
CATCH：ICLR 2025 最值得关注的时间序列异常检测新框架
Anomaly Detection in Time Series
面向零样本时间序列异常检测的基础模型：利用合成数据和相对上下文差异
Programming
Thoughtworks Technology Radar 33
vexor - vector-powered CLI for semantic search over files
mgrep - A calm, CLI-native way to semantically grep everything, like code, images, pdfs and more
Google Brings Colab Integration to Visual Studio Code
DuckDB Extensions：让本地分析更强大的秘密武器
System Design Interview: Design Twitter/X Timeline - A Frontend Deep Dive
AWS Lambda Rust Support Reaches General Availability
Other
I analyzed 1000 forward deployed engineering jobs – what I learned
史诗级Bug！OpenReview全员裸奔，匿名评审秒变「实名大逃杀」