One idea leads to another
Stay with the thread.
Ordered explanations and recurring formats. Counts reflect published articles; planned collections are labeled clearly.
Inside AI Systems — One Problem, One Architecture, One Lesson. ↗
A daily educational reading series: investigate one practical problem, trace an architecture, and test what you learned. Start with seven complete lessons about an AI student assistant.
Machine Learning from First Principles ↗
Understand what models learn, how they fail, and how to measure them.
Deep Learning Under the Hood ↗
Follow information and gradients through a neural network.
NLP to LLMs ↗
From text representation to language models.
The Transformer Series ↗
Representation, attention, and the transformer architecture.
Generative AI Systems ↗
Connect models with context, retrieval, and evaluation.
The Agentic AI Series ↗
From a model response to a controlled loop of actions.
The Architecture Series ↗
See the boundaries, flows, and trade-offs of AI systems.
AI in 5 Minutes ↗
One useful mental model at a time.
Under the Hood ↗
Trace the mechanisms behind the interface.
Paper → Product ↗
Read the evidence, then consider the engineering implications.
Model Shift ↗
Evidence-led analysis of meaningful model developments.
Build This ↗
Small, inspectable implementations to learn by doing.
One Big Idea ↗
A question that changes how you frame a system.