Technical Blog
Notes on data engineering, Spark, Kafka, and cloud architecture.
DuckDB: Big Analytics, Small Footprint
Several GB, one Python process: how DuckDB works, why vectorization and compression matter, and when to choose pandas, Polars, Spark, Snowflake, or Databricks instead.
How to Best Use AI Coding Agents Like Claude
The climb from 'fancy autocomplete' to orchestrating a team of agents in parallel — a six-level guide covering briefing, CLAUDE.md memory, Skills, subagents, MCP connectors, and parallel worktree workflows.
Disabling Default Outbound Connectivity in Azure
An introductory security guide to Azure default outbound access, explicit egress, and choosing a controlled internet path.
Building a Governed Multi-Agent FinOps Platform
An engineering architecture for turning trusted cloud cost data into governed decisions using modular services, time-series storage, specialized AI agents, and human approval.
Compression Algorithms You Probably Inherited: gzip, Snappy, LZ4, zstd — and What’s Newer
A practical, benchmark-backed comparison of gzip, Snappy, LZ4, zstd, Brotli, and XZ across public text, JSON, mixed-file, and audio datasets.