Shared services work management
A shared work management platform for 90+ finance staff: current work items, consistent statuses, persistent notes, and automated reporting in one place.
Read the case study: Shared services work managementWork
Systems that make the work easier: shared platforms, clearer decisions, and less manual coordination. Explore the problem, the approach, and what changed.
Direct API ingestion framework
A reusable Python and Snowflake framework that loads marketing and commerce platform APIs directly into the warehouse, with shared logging and append-only landing tables, replacing third-party connectors for major sources.
Read the case study: Direct API ingestion frameworkMarketing measurement models
The spend, sales, and attribution models behind marketing mix modeling, CAC reporting, and multi-touch attribution for the organization, with QA monitors and governed metric definitions - and an honest boundary around what is still in development.
Read the case study: Marketing measurement modelsData modeling and lineage framework
An in-house SQL templating and modeling framework for AWS Athena with column-level lineage, dependency graphs, and generated BI documentation - and the later decision to migrate to SQLMesh.
Read the case study: Data modeling and lineage frameworkUnified orchestration engine
A Rust job scheduler that replaced orchestration fragmented across Airflow, AWS Step Functions, and Dagster with one operational surface: shared scheduling, error handling, and operational visibility.
Read the case study: Unified orchestration engineInventory prediction, and the courage to delete it
A clustered regression engine for advertising inventory scenarios - evaluated honestly, found no better than a simple baseline, and replaced by a query-based approach with fewer anomalies and simpler extensions.
Read the case study: Inventory prediction, and the courage to delete itPipeline automation at API-limit scale
Automating management of thousands of Datorama vendor pipelines for the organization with a Python API wrapper: cached configuration, adaptive job checking, and specific detection, retry, and escalation behavior instead of manual queue-watching.
Read the case study: Pipeline automation at API-limit scaleLegacy deployment automation
Giving Informatica a real CI/CD path for the organization: Logic Apps and GitHub Actions driving the Informatica API to promote objects between environments, with automated remapping of the environment-specific IDs that made API-driven promotion hard.
Read the case study: Legacy deployment automation