Work

Production systems. Live demos.

Selected work spanning ML, automation, and full-stack apps. Many engagements are under NDA — these are the ones that can be shown.

Live demo

Interactive BI Dashboard

Live BI Demo · Mock data

Marketing & Forecast Dashboard

Revenue
$1.02M
+5.3%
Conversion
3.15%
+0.4%
Active Customers
12,408
+182
Forecast (next 24h)
$41.1K
model: BR-v4
Revenue · Actual vs Forecast
tick #0
Channel Mix
Conversion Rate · 7-day rolling
Avg 3.15%

Demo data. The same patterns ship in Power BI, Tableau, Looker, and custom Next.js dashboards.

Case studies

Real systems, real outcomes

ML / RUNNING 9 YEARS

ZIP-Code Lead Prioritization

Tobit regression on American Community Survey data — built for a home-services call center to prioritize incoming leads when granular data wasn't available. Still in production nine years later. Same algorithm. Same model.

Years in production
9
Industry
Home services
Stack
Python · Tobit regression · ACS
ML / COMMODITY FORECASTING

Avocado Demand Forecasting

Bayesian regression on a real-world commodity. Beat the open market three years running for one of the largest US avocado suppliers.

Annual commodity volume
$100M+
Years beating the market
3
Stack
Python · Bayesian regression
ML / OPTIMIZATION

$300M Bidding Engine

Designed and shipped a recommendation + bidding optimization engine for a national auction platform. Real money, real-time bidding, real stakes.

Annual GMV
$300M+
Role
Lead ML / engineer-of-record
Stack
Python · GBM · real-time bidding
AUTOMATION / ENTERPRISE

Reporting Stack Overhaul

Took a multi-team operation from 20+ source spreadsheets to a single automated pipeline feeding Power BI, Tableau, and Looker — in roughly 20 hours of work.

Sources unified
20+
Time to ship
~20 hrs
Stack
dbt · Power BI · Tableau · Looker
Live apps shipped

Things you can click on right now

Got a system that needs to ship?

Book a 15-minute call. By the end of it, you'll know exactly what it costs and how long it takes — no fluff, no "discovery phase."