All lessons

Jun 18, 2026·3 min read

Understanding Transformers from first principles

A clear-eyed look at the mechanics, the trade-offs, and the parts most write-ups quietly skip over.

#llm#transformers#deep-dive

Jun 15, 2026·3 min read

A practical guide to Diffusion Models

We trace the idea from a napkin sketch to a system that holds up under real production traffic.

#vision#generative

Jun 12, 2026·3 min read

RLHF: what actually works in production

Less theory, more of the messy decisions you actually face when shipping this into the world.

#rlhf#safety#training

Jun 9, 2026·3 min read

Notes on Retrieval-Augmented Generation, without the hype

What the benchmarks tell you, what they hide, and how to read the difference between them.

#rag#retrieval#nlp

Jun 6, 2026·3 min read

Inside Agentic Workflows

A field guide for engineers who would rather understand the why than memorize the how

#agents#tools#infra

Jun 3, 2026·3 min read

The quiet trade-offs of Quantization

A clear-eyed look at the mechanics, the trade-offs, and the parts most write-ups quietly skip over.

#serving#infra#eval

May 31, 2026·3 min read

Why Embeddings keep surprising us

We trace the idea from a napkin sketch to a system that holds up under real production traffic.

#embeddings#nlp#retrieval

May 28, 2026·3 min read

Model Evaluation: what actually works in production

Less theory, more of the messy decisions you actually face when shipping this into the world.

#eval#serving

May 25, 2026·3 min read

Inside Inference Optimization

A field guide for engineers who would rather understand why than memorize how.

#serving#infra