Notes on dataarchitecture.
Short essays on architecture, cost, hiring and AI: one cartoon, one idea, most days of the week. No tutorials, no listicles.
The Stack You Actually Need At 15 Engineers
At 15 engineers, a data stack with twelve tools is mostly twelve invoices and one tired person.
Read →What a Broken Pipeline Actually Costs Per Week
Pull the last four weeks of pipeline incidents and put an hour next to each one. The total is rarely what you guessed.
Read →Coalesce Context As Infra
Treat your metric definitions like a Terraform file, or keep the agent away from them.
Read →DE Role Evolution 2026
If a model writes the pipelines, I don't know yet where your juniors learn why pipelines break.
Read →Semantic For Unstructured First
Most AI data problems I see start in a column called payload, long before any embedding.
Read →Orchestration Consolidation Matrix
Count the places a job can be scheduled in your data stack. Include the cron on that one VM.
Read →BigQuery Scan Cost Surprises
LIMIT 10 doesn't reduce what BigQuery scans on an unclustered table. A lot of teams learn that from the invoice.
Read →dbt Coalesce Metadata Shift
The description fields in your dbt project used to be read by nobody. Now an agent reads them.
Read →Governance Before AI Act Fines
The provenance record behind an AI feature fits in a YAML file next to the pipeline code.
Read →Process Before People Fix
Before you replace the data engineer who seems slow, count how many people can hand them work.
Read →Want expert eyes on your data architecture?
No pitch. An honest conversation about whether I can help, and what shape it would take if I can.