Portfolio

Projects

Each project below is something I designed and built. They cover the full spectrum of modern data engineering: cloud lakehouses, open-source pipelines, entity resolution, AI-powered automation, and developer tooling. Most are live and running in production.

Azure Databricks Lakehouse

DATABRICKS

JDE Aerospace ERP on Databricks

End-to-end data lakehouse for JD Edwards aerospace data. Full medallion architecture with Lakeflow Declarative Pipelines (DLT), SCD Type 2 via APPLY CHANGES INTO, DLT Expectations for data quality, and Unity Catalog governance with role-based access control. Achieved 96% Photon engine execution with zero disk spill. Includes AI/BI Dashboards, Genie Spaces, and a multi-task Workflow DAG.

Azure DatabricksLakeflow DLTUnity CatalogDelta LakeSpark SQLPySparkPhotonAI/BI Dashboards

JDE Data Platform

LIVE

Open-source medallion architecture

Full production data platform built on open-source tools. Data flows from a simulated JD Edwards ERP through Bronze/Silver/Gold layers into live dashboards, status boards, a customer self-service portal, and an e-commerce storefront. Node.js extractors pull from JDE SQL Server into PostgreSQL via Apache Airflow orchestration. dbt Core handles transformation with Kimball dimensional modeling.

PostgreSQLdbt CoreApache AirflowNode.jsNext.jsFastifyRailway

MDM Entity Resolution

MDM

Cross-ERP golden records with Splink

Master Data Management layer that unifies customer records across five separate ERP systems, each with their own schemas, naming conventions, and duplicates. Uses Splink probabilistic record linkage to resolve entities like "Boeing Co.", "THE BOEING COMPANY", and "Boeing Defence UK Ltd" into golden records. Resolved 82 duplicates from 333 source records into 251 golden customer entities with consolidated sales visibility.

PythonSplinkPostgreSQLDuckDBDatabricks ServerlessJaro-WinklerFellegi-Sunter

SEC Filing Analyzer

AI

LLM-powered financial document analysis

Automated pipeline that pulls SEC EDGAR filings, processes them through Claude API for intelligent summarization and structuring, and stores results in PostgreSQL. Orchestrated with n8n for scheduled pulls and event-driven processing. Designed for analysts who need rapid, structured insight from public company filings without manual review.

n8nClaude APIPostgreSQLSEC EDGARLLMPython

SCD2 Change Tracker

AI

Intelligent change detection and alerting

SCD Type 2 change detection system for JDE master data. PostgreSQL triggers capture every change to critical business entities, n8n orchestrates the alerting workflow, and Claude AI generates intelligent summaries of what changed and why it matters. Designed for data stewards who need to know when master data shifts without watching it manually.

PostgreSQLn8nClaude AIPL/pgSQLSCD Type 2Webhooks

SAP Medallion Engine

TOOL

Code generator for Databricks/Kimball pipelines

Python/Jinja2/YAML code generator that produces Databricks pipeline code from YAML configuration files. Encodes Kimball dimensional modeling standards, entity-first naming conventions, and current Databricks best practices including CLUSTER BY AUTO, Lakeflow Declarative Pipelines, and AUTO CDC. Built to accelerate pipeline development for ERP data migrations.

PythonJinja2YAMLDatabricksLakeflow DLTKimballDAB