Most of my work is rebuilding broken data foundations, usually a migration off legacy BI, sometimes a fresh build. The destination is a modern, AI-ready stack, sized for your scale: dbt + Snowflake + Fivetran + Omni at the high end, leaner alternatives for earlier-stage brands. The before-state is usually fragmented data, dashboards stakeholders don’t trust, and AI that can’t be reliably used for self-service analytics. The after-state is models and metrics everyone can rely on: stakeholders, investors, and AI.
Defining the metrics that matter — cohort retention, LTV, conversion, health outcomes — and getting them agreed-on across stakeholders. The hardest part of the job, and the one a clean stack is meant to deliver on.
Snowflake warehouses provisioned to scale with your growth — one platform for all your data, no silos to maintain.
Centralized data from your operational tools — payments, e-commerce, CRM, marketing — into Snowflake using Fivetran. Everything in one place, ready to model.
Fast, well-documented dbt models your team can extend. Or migration and cleanup of an existing dbt project that’s grown into spaghetti.
Legacy BI (Tableau, Looker, Mode) migrated to a modern, AI-ready stack — dbt + a semantic layer like Omni. Reports are fast, metrics stay consistent, and your data is ready for AI.
Subscription supplement, clinical-data, and longevity brands. Cohort retention, biomarker timelines, member-journey models — the data foundation that supports investor-grade outcome reporting.
Multi-channel commerce — Shopify, Amazon, retail. Cross-channel unit economics, repeat-purchase modeling, channel-mix margin analysis. Including subscription unit economics for recurring-revenue brands.
Replacing legacy BI (Tableau, Looker) with modern, AI-ready stacks.