Planet of the Vapes.

The project

An e-commerce business depends on many systems agreeing: the storefront, inventory, warehouse, carrier, customer service, marketing, and accounting. My work connects those systems and gives people a shared way to understand the business.

My contribution

I took over the data warehouse from a three-person team and built tools that bring operational data, contribution-margin analysis, and agent-assisted work together. I work across React and TypeScript interfaces, Python services, Databricks, and AWS infrastructure.

Project details

• Connected Shopify, Cin7, and ShipNetwork through webhook consumers and scheduled synchronization, with retries and duplicate protection for order fulfillment.
• Integrated Shopify with Klaviyo, Gorgias, and Xero, and migrated n8n automation to AWS Lambda and infrastructure defined in AWS CDK.
• Built internal Java and Node.js/TypeScript services, and configured a multi-account AWS organization with IAM and service control policies.

A warehouse teams can work from

Deployed recurring ingestion jobs with Databricks Asset Bundles, moving Shopify, support, web analytics, and SEO data from Amazon S3 into bronze and silver Delta Lake tables. Built Spark SQL pipelines for order and inventory analysis, reusable Python notebooks with approved calculations, and a SQL Warehouse for querying the catalog.

Ask a question, inspect the analysis

Deployed a Streamlit chat interface on Databricks Apps with AI/BI Genie and custom FastMCP tools for contribution-margin analysis. A retrieval workflow supplied approved business definitions. Service roles and controlled execution environments limited access while agents produced charts and dashboards.

Share the result

Built a React and TypeScript collaboration portal on AWS for agent-generated charts and reports. Moved model inference from Databricks Model Serving to Amazon Bedrock to reduce hosting costs, and worked with marketing and support on GoHighLevel community workflows.