The Index

Topics.

Explore the ideas, systems, and engineering problems behind modern ML infrastructure.

01

Inference & Serving

vLLM, TGI, paged attention, continuous batching, speculative decoding. 4 articles 02

Training Systems

Trainers, optimizers, recipes, debugging large runs. 1 article 03

Architecture

Transformers, MoE, SSMs, hybrids, and what's next. 4 articles 04

Distributed Training

FSDP, tensor parallel, pipeline parallel, sequence parallel. 1 article 05

Quantization

PTQ, QAT, FP4, FP8, mixed precision, calibration. 1 article 06

Retrieval & RAG

Embeddings, indexes, re-rankers, and pipeline systems. 2 articles 07

Models

LLMs, VLMs, multimodal systems, capabilities, and model behavior. 1 article 08

Agents

Planning, tool use, multi-agent systems, memory, and orchestration. 1 article 09

Evaluation

Benchmarks, harnesses, contamination, signal vs noise. 1 article