More than twenty years running large web and email platforms across Amazon, Rosetta Stone, and James Madison University, turning platform work into measured outcomes and building the tools, in Python, Ruby, and AI agents, that remove the manual work.
More than twenty years across Amazon, Rosetta Stone, and James Madison University turned platform work into measured outcomes. Starting as an Amazon software engineer built a root-cause-first, customer-first habit that carried into every role after. Today that means owning quality, uptime, and cost for a 34,000 page platform serving more than 500 content owners.
At JMU, retiring two paid platforms by building their replacements in-house eliminated roughly $90,000 a year in vendor and manual cost. Redesigning three strategic pages around request-information and schedule-a-visit lifted prospective-student conversions 42 percent. Minute-level homepage monitoring and a daily 404 report have sustained availability with no downtime in 12 years, surfacing issues before the system administrators.
Recent work leans on AI directly. Prototyping design directions with AI turns ideas into reviewable mockups quickly, and an agentic content-QC pipeline keeps the platform clean: Python watchers flag new or changed pages and capture a screenshot of each change for visual review, then a daily LLM agent proposes editorial fixes, SEO title improvements, and content tags for human approval.
Every filled cell is a month held up on a 34,000 page platform, twelve years without downtime and counting into 2026. Brighter cells mark the year a paid tool was retired, a platform was rebuilt in-house, or a milestone shipped. Select a year to see what shipped.
Hover or tap a year column to see the platform milestone that shipped that year.
Owning quality, uptime, and cost for a 34,000 page platform serving more than 500 content owners, retiring paid tools by building in-house replacements, and governing the CMS to 71 canonical object types.
Coordinating online and mobile product delivery for teams of up to 45 across engineering, QA, and stakeholders, and shipping the Rosetta Stone Kids product.
Leading a 5-person engineering team delivering 13 production services and 24 libraries, and replacing Adobe LiveCycle Collaboration with an in-house build in 9 months.
Building the org's first developer training program and owning 42 code packages across 11,863 files, earning Outstanding, the highest rating.