Technical perspectives on cloud infrastructure, modern data platforms, and the architecture decisions that matter.
Architecture write-ups and field notes, updated as the thinking evolves.
A D2C marketing stack simulated from a known ground-truth process, centralised through dbt, and measured with a Bayesian marketing-mix model — where last-click's channel credit is provably wrong by construction, and every corrected number is derived rather than asserted.
Read — ↗A supply-chain lakehouse built on Databricks and the medallion architecture — from internally consistent simulated data through Bronze and Silver refinement to Gold decisions, with every headline number derived rather than asserted.
Read — ↗An end-to-end data architecture built with AWS, Snowflake, and dbt, structured around the medallion architecture — moving data from raw ingestion to trusted, business-ready outputs in a way that is secure, maintainable, and ready to scale.
Read — ↗A technical comparison of object, block, file, and analytical storage across the three major cloud providers — when to use each, and the trade-offs that matter.
Read — ↗