- LLM Inference Became A Systems Problem
How batching, caching, quantization, and speculative decoding changed serving economics.
8 min - Agentic Workflows Need State And Guardrails
Why useful LLM agents depend on tools, state machines, evals, and careful failure handling.
7 min - Multimodal LLMs Became Product Infrastructure
How vision, audio, and text models changed document workflows, support, and automation.
5 min - Small Language Models Found Their Lane
Why smaller LLMs became useful for routing, extraction, classification, and edge workflows.
5 min - Long Context Is Not A Retrieval Strategy
Longer context windows help, but they do not replace retrieval, ranking, and context design.
7 min - Structured Outputs Made LLMs Easier To Ship
Why schemas, tool calls, and constrained decoding changed how production LLM apps are built.
6 min - LLM Evals Became the New Unit Test
Why evaluation became a core skill for building reliable LLM applications.
9 min - QLoRA Made Fine-Tuning Feel Practical
How LoRA and QLoRA changed the economics of adapting LLMs.
8 min