Data pipelines, warehouses, BI, ML, and AI engineering.
The lakehouse pattern is replacing pure warehouse architectures at scale. The two open table formats — Apache Iceberg and Delta Lake — and how to choose.
Vector DBs were the hot category of 2023. Most teams that adopted one did not need it. Here is when you genuinely do.
LLMs are now load-bearing in real applications. The reliability engineering practices for keeping them up — eval harnesses, fallbacks, observability.
Most enterprises have dashboards nobody opens. The cultural changes that make data actually drive decisions.
Data warehouses became the source of truth. Then ops teams needed that truth back in the apps they used. Reverse ETL was inevitable.
Three warehouses, three different sweet spots. The decision criteria that actually matter at scale.
A dbt codebase eventually becomes a software project. The patterns that work at 5,000 models — and the ones that fall apart.
The "modern data stack" promise five years ago was a clean toolkit for analytics. The reality in 2026 is messier and more interesting.