Posts
All the articles I've posted.
- 8 MIN READ•May 28, 2026
The 2026 Unified Data Architecture: Reconciling Multi-Cloud Data Lakehouses
Multi-cloud data lakehouses in 2026 run on Apache Iceberg, open catalogs, and zero-ETL federation. Here's what a composable, unified architecture looks like.
Unified Data Architecture 2026 - 6 MIN READ•May 28, 2026
Why Traditional Lakehouses Fail AI Agents: The Mathematical Case for the Agentic Lakehouse
Traditional lakehouses expose raw directories and ambiguous schemas to AI agents, causing hallucination. Here's the mathematical case for why this fails and what fixes it.
Why Lakehouses Fail AI Agents - 7 MIN READ•May 28, 2026
The Era of Zero-ETL Federation: Fueling AI Agents with Real-Time Cross-Enterprise Data
Zero-ETL federation lets AI agents join real-time CRM data with historical lakehouse tables instantly. Learn the architecture, tradeoffs, and how Dremio enables it.
Zero Etl Federation AI Agents - 7 MIN READ•May 25, 2026
Use Hermes Agent for Free With DeepSeek V4 and Slack
Hermes Agent is a free, open-source AI agent from Nous Research. Connect it to DeepSeek V4 for zero-cost inference and Slack for anywhere access. Here is how to set it up in 10 minutes.
AI ToolsOpen SourceHermes Agent - 13 MIN READ•May 24, 2026
Automating Table Maintenance Before Small Files Accumulate
Learn how Databricks Predictive Optimization, AWS S3 Tables, and Iceberg native actions automate compaction and snapshot management before small files degrade performance.
Iceberg Table Maintenance AutomationIceberg CompactionSmall Files Lakehouse - 16 MIN READ•May 24, 2026
Choosing the Right Iceberg Control Plane: Polaris vs. Unity Catalog vs. Cloud REST
Choosing an Apache Iceberg catalog? Compare open-source Apache Polaris, open Unity Catalog, and managed cloud REST control planes to unify your lakehouse.
Apache Iceberg CatalogIceberg Rest CatalogApache Polaris - 13 MIN READ•May 24, 2026
Clean Rooms for Privacy-Preserving Analytics
Data clean rooms enable secure multi-party analytics without sharing raw data. Learn how Databricks Clean Rooms, AWS Clean Rooms, and BigQuery differential privacy work.
Data Clean Rooms Privacy-Preserving AnalyticsDatabricks Clean RoomsAws Clean Rooms - 14 MIN READ•May 24, 2026
Building Composable Query Engines with Rust Runtimes
Apache DataFusion, Velox, and Substrait form the foundation of modern composable query engine stacks. Learn how these components fit together and when to use each.
Composable Query Engine DatafusionApache Datafusion RustVelox C++ Engine - 11 MIN READ•May 24, 2026
Data Mesh After the Hype: What Actually Works
Three years after Zhamak Dehghani's original papers, data mesh has proven valuable in specific organizational contexts and impractical in others. Here's what the practical implementations look like.
Data Mesh Practical ImplementationData Mesh Reality CheckData Product Thinking - 12 MIN READ•May 24, 2026
How dbt Fusion Reshapes Analytics Engineering
dbt Fusion entered public beta in May 2025 with a Rust-powered runtime that changes how analytics engineers develop, validate, and deploy SQL models. Here's what changed.
Dbt Fusion Analytics EngineeringDbt Fusion RustDbt State-Aware Orchestration - 13 MIN READ•May 24, 2026
Using DuckDB and Polars to Query Iceberg Tables
DuckDB 1.4 LTS and Polars streaming engine now both support reading and writing Apache Iceberg tables. Learn how to use them for lakehouse analytics in 2025.
Duckdb Polars IcebergDuckdb Iceberg WritePolars Iceberg Sink - 12 MIN READ•May 24, 2026
FinOps for Data Warehouses with Open Billing Data
The FOCUS 1.3 specification and native warehouse cost views make real-time cost attribution practical. Learn how to build a FinOps pipeline for Snowflake, BigQuery, and multi-cloud environments.
Warehouse Finops Focus SpecificationSnowflake Cost ManagementBigquery Jobs View