Projects

Real builds, real constraints, real outcomes.

Across data platforms, analytics, and AI, in media, retail, government, and automotive.

Enterprise work

Delivered in production, at scale

Data and analytics systems built inside a large consumer business, anonymized for confidentiality.

Data Engineering

Enterprise Data Foundation

One trusted model, feature store, and governed semantic layer across five source systems, so BI, ML, and AI build on the same numbers instead of fighting over them.

5

Source Systems Unified

Millions

Profiles Modeled

Databricks Delta Lake SSAS Power BI
BI & Analytics

Enterprise Semantic Layer & KPI Framework

One governed definition of every KPI, so teams stopped reconciling conflicting numbers and report builds dropped 60-70%.

50% Faster

Query Performance

100+

Governed Measures

SSAS DAX Power BI SSIS
AI & Automation

Voice-of-Customer Intelligence Platform

Feedback from four social sources unified into one searchable layer that anyone can question in plain language, turning hours of analyst work into seconds.

4

Social Sources Unified

Hours to seconds

Sentiment Access Time

Databricks Vector Search LangGraph Genie
AI & Automation

CRM Campaign Automation Platform

Daily CRM campaigns that build and target themselves across millions of profiles, replacing multi-day audience handoffs with hands-off execution.

Millions

Profiles Processed Daily

4

Recommendation Scenarios

Power BI Databricks PySpark
Data Engineering

Ad Inventory & Revenue Pipeline

Pacing shortfalls and inventory pressure surface during the campaign, in time to act, instead of reconciling make-goods after it ends.

Same-day

Pacing Visibility

Multi-week

Settlement Refresh

Databricks S3 Power BI
Data Science

Behavioral Segmentation & Attribute Inference

Raw behavior turned into governed, activatable segments and inferred attributes, so marketing and planning act on what customers do instead of sparse signup data.

Millions

Profiles Scored

~4x

Attribute Coverage

scikit-learn XGBoost MLflow Databricks

Personal builds

Built end to end, on my own

Full-stack systems I designed, built, and deployed myself, from ingestion to interface.