Drop Taavi into Claude and get a semantic layer engineer who reconciles the three conflicting revenue numbers first, then makes one definition the only one anybody can query. Taavi owns the single definition of every business metric: semantic layer implementation across the dbt Semantic Layer and MetricFlow, Cube and AtScale; metric specification covering grain, filters, time dimensions, additivity and non-additive measures; dimensional consistency and slowly changing dimensions in metric context; metric versioning and deprecation; headless BI serving the same metric to every downstream tool; governance of who may change a definition; reconciling conflicting numbers between teams step by step until the ladder closes; query pushdown and caching; and the migration path from a pile of hardcoded dashboard SQL to governed metrics. What you get →Metric specs: grain, filters, time dimensions, additivity →Reconciliation ladders that close between conflicting numbers →dbt Semantic Layer, MetricFlow and Cube implementation →Metric versioning, deprecation and change control 📄 taavi-metrics-semantic-layer-engineer.skill Under 2 min install Works with Claude, ChatGPT & any AI chat How to install Download the .skill package → open Claude → paste SKILL.md into your Project Instructions or system prompt → describe your requirement → Taavi builds the answer. Includes a full worked example so you see exactly what you get.
