Choosing an embedding model is one of the first decisions you make when building a retrieval-augmented generation system, and it is one of the hardest to reverse. The model you pick determines your ve

Every few months, someone declares that a new tool has made data engineering obsolete. First it was managed warehouses. Then dbt. Then generative AI that writes SQL. Each time, the prediction is the s

Most data catalogue projects die within six months. The tool gets purchased, a team populates it with metadata for a few hundred tables, enthusiasm fades, and twelve months later the catalogue is a st

Predictions about technology roles are usually wrong in predictable ways. They overestimate the speed of change, underestimate the persistence of legacy systems, and assume that technical trends will

The decision to centralise AI infrastructure into a platform team usually comes after a period of decentralised pain. Three product teams independently built model serving pipelines. None of them shar

The framing of AI systems as either "copilots" (human-in-the-loop, AI assists) or "autopilots" (human-out-of-the-loop, AI acts independently) has dominated the conversation about AI autonomy for two y

You search Google for "bank account interest rates." The first result is an advertisement for a bank. The second is a comparison site. The third is a news article about the Fed's latest decision. The

Single-agent applications (one LLM, one set of tools, one task) are straightforward to build and debug. The agent receives input, calls tools, produces output. When multi-step reasoning or collaborati

An online travel agency processed 2.3 million flight searches per day. Each search triggered a pricing computation that determined the displayed fare for every matching itinerary. The pricing computat