Lakehouse Context Layers with Atlan and Iceberg v3
The context layer explains what lakehouse data means, which is the part table formats do not solve alone. That is the useful lens for…
The archive
Writing, page 19
593 posts on Apache Iceberg, lakehouse architecture, data engineering and applied AI.
Real-Time Agentic Analytics with ClickHouse
Real-time agents need analytical systems that can answer while an event still matters. That is the useful lens for real-time agentic…
Composable Analytics Beats Metric Catalogs
Metric catalogs tell agents what terms mean. Composable analytics tells agents how to reason with those terms safely. That is the useful…
Goal-Directed Analytics Agents on Apache Iceberg
The next step after text-to-SQL is a governed action loop with checks before every external effect. That is the useful lens for…
Iceberg Remote Signing for Regulated Datasets
Remote signing is the stricter pattern for lakehouse storage security because clients request signed file operations instead of receiving…
Apache Iceberg v3 Deletion Vectors on Snowflake
Deletion vectors matter because row-level changes should not require a full rewrite of every affected data file. That is the useful lens…
CDC Without Complexity Using Iceberg v3 Row Lineage
Row lineage gives Iceberg a native way to tell incremental consumers which rows changed and when they changed. That is the useful lens for…
The 2026 Guide to Iceberg View Federation
Portable views are the missing logic layer between open tables and multi-engine analytics. That is the useful lens for Iceberg view…
Implementing MCP in the Lakehouse
MCP gives AI clients a standard way to call governed lakehouse tools instead of guessing how to query your data. That is the useful lens…
Microsoft Fabric Build 2026 Agentic Analytics Stack
Microsoft's Fabric direction shows that agentic analytics is becoming a platform architecture, not a chat feature. That is the useful lens…