About Signal & Syntax

Most writing about LLMs sits at one of two altitudes: high-level explainers aimed at executives and curious newcomers, or dense research papers aimed at other researchers. There’s less writing in the middle, where working engineers actually operate. Most engineers need to understand enough of what’s happening under the hood to build and debug real systems, but not so much that they feel like they’re replicating a PhD program.

Signal & Syntax serves that middle altitude. Posts work through the mathematics and engineering underneath the models we’re all building on: gradient descent, attention, tokenization, inference-time behavior, training infrastructure, evaluation, orchestration. The goal is to write about this material the way one engineer explains something to another, with enough rigor to be trustworthy, enough plain English to be readable, and enough specificity that the post helps you build something.

What You’ll Find Here

• Under the hood: how components actually work. Attention mechanisms, tokenizer behavior, loss landscapes, numerical stability.
• Applied engineering: how working systems are built. Orchestration patterns, evaluation frameworks, deployment considerations, production tradeoffs.
• The mathematics: linear algebra, probability, and optimization — the foundations that make everything else make sense.
• Occasional essays: reflections on the industry, the technology, and what it’s like to build in this moment.

New Here?

Start with How LLMs Read Code for a representative post, or browse the Categories and Tags in the sidebar.