Enterprise AI infrastructure, automation, cloud-native architecture, and data solutions.
Two-tier, metadata-first design for permission-aware retrieval in enterprise AI assistants.
Role: Senior Technical Program Manager
An AI-powered enterprise feature-request and roadmap platform, built end-to-end.
Role: Sole designer and builder
A serverless system that summarizes and analyzes project data daily, no human in the loop, ~1 year in production.
Role: Sole technical owner
Interactive "learn by doing" lessons that turn non-engineers into producers of AI-generated work.
A framework that disaggregates AI usage by workforce segment, showing who actually benefits.
Role: Owner and builder
A personal-finance SaaS that projects your balance a year ahead and generates payment strategies. In beta at bills-app.com.
Role: Sole designer, builder, and operator
A collections-management platform that ran a real Chilean collections operation for about a year, managing thousands of debt records through a 22-state workflow.
A collaborative fraud-screening MVP with fuzzy identity matching and a documented traffic-light REST API. The product was complete; the venture closed when go-to-market stalled.
A self-updating gallery of NASA's Astronomy Picture of the Day, fetched daily by Lambda and archived to S3. Runs for under $0.01/month.
A real-time dashboard for e-commerce order picking across multiple store locations, with automatic alerting for at-risk deliveries.
A centralized inventory lookup for customer service reps, cutting response times from 15+ minutes to 2 by unifying VTEX, SAP, and in-store stock data.
Turned shelf-scanning robot data into exact product locations (aisle + shelf coordinates) across 16 supermarket stores, searchable from the sales floor.
A Python tool that analyzes field usage on any Salesforce object, helping admins find and retire underutilized fields. Open source on GitHub.
A system designed to compute real-time on-shelf stock from SAP and sales data, telling staff what to restock, how much, and where, then reusing shelf-location data to plan optimal picking routes.