Projects

Designed and delivered an end-to-end, AI-native intelligence system for a global services organization. A data-confidence layer validates and scores inputs before any AI touches them; a governed business-insights layer turns that data into action-oriented executive pages with AI-summarized narrative and clear recommendations; and an automated operating-review pipeline feeds forecasting and planning. Adopted organization-wide as the standard for executive reporting — replacing “whose data is right?” debates with one trusted, action-first view.

AWS Global Services had field data everywhere and insights nowhere. I was handed a whiteboard and told to fix it. What came back was the Insights Hub — AWS’s first centralized analytics platform for the field — along with the Insights Catalog, a bi-directional data framework, and a Data Confidence and Completeness Scoring system that made the data trustworthy enough to build on. The AI-ready data layer became the foundation for predictive modeling and GenAI integrations across the org.

Bloomberg ran on infrastructure nobody could see clearly. Hundreds of operational systems held pieces of the picture — CMDBs, ticketing platforms, monitoring tools — but nothing reconciled them into a single source of truth. I built that system from scratch: a unified platform that ingested, reconciled, and surfaced the full lifecycle of Bloomberg’s infrastructure estate. The engineering org ran on it. Nothing like it had existed before.

Leadership can’t run a global business on gut feel. I designed the KPI architecture, data dictionaries, and WBR/MBR/QBR review mechanisms that AWS Global Services used to manage performance — from the field up to the C-suite. When the weekly business review ran, the data it ran on was mine.

The global TAM organization had no systematic way to see which customers were healthy and which were at risk. I built one. Starting with the analytics team standup, I engineered the first customer health scoring and churn-risk models the field had ever had — giving leadership a data-driven view of renewal and risk that changed how account teams prioritized their time.

Replaced a slow, manual operating-review process with a serverless, AI-driven pipeline: it gathers program signals from work-management and knowledge systems, uses AI to compose a structured executive narrative, and generates and distributes the finished report automatically. Built as infrastructure-as-code with least-privilege security and full observability. Shipped quickly and adopted across the organization — a template for turning any recurring reporting grind into an automated, governed workflow.

Bloomberg’s Windows estate ran on legacy provisioning that couldn’t keep pace with the organization’s growth. I led the Active Directory modernization and the global virtualization program — heading a 30+ engineer team that replaced manual processes with automated, scalable infrastructure across Bloomberg’s worldwide environment. 24/7, 365. No exceptions.

Architected and executed the migration of the Bloomberg Terminal from legacy proprietary systems to commodity Linux at enterprise scale — and built the firm’s first private-cloud platform along the way — with zero production disruption on a mission-critical product used worldwide. Automated global provisioning enabled rapid, consistent deployment. If you use the Terminal today, you’re running on this infrastructure.

When I took over Global Linux Engineering, Bloomberg provisioned servers by hand. I built the private cloud that changed that — architecting the automation layer, standing up the CMDB and telemetry platforms, and scaling the Linux estate from a small footprint to tens of thousands of hosts globally. Bloomberg is 24/7, 365. The Terminal never goes down. By the time I moved on, manual provisioning was a memory.

The Daily News ran presses through the night, every night. Editorial closed, the building quieted, and production took over — plates made, ink running, papers out the door before the city woke up. I owned the full IT estate that kept it going: servers, networking, the PLC systems that controlled the production floor, and the AGV infrastructure that moved materials through the plant. Overnight production meant overnight problems. Something goes down in the middle of a press run, you fix it. No ticket queue. No morning. Six years of that builds an operational instinct that never leaves you.

Stood up the business-operations backbone for one of the largest global services businesses in cloud — from a small founding team into a multidisciplinary global function on a hub-and-spoke model. Built the capacity-planning mechanism, the KPI scorecards, the annual operating-plan cycle, and the weekly/monthly/quarterly executive review cadence that leadership ran the business on. The result: growth and efficiency that didn’t depend on heroics.

The TAM teams covering Financial Services, Healthcare, Media & Entertainment, and Government had no unified view of their business. I built the revenue-operations framework, the P&L dashboards, and the analytics infrastructure that changed that — giving field leadership real-time visibility into performance, churn risk, and renewal health for the first time.

Enterprise Support was growing faster than the operating model could keep up. I built the structure that caught up with the scale — a hub-and-spoke segmentation model, customer prioritization and tiering frameworks, and a multidisciplinary team assembled from scratch: PMs, business intelligence engineers, data scientists, TPMs, and curriculum developers. I stood up the function and defined how it ran.

Amazon’s Bar Raiser program is an invitation you earn, not apply for. It recognizes the handful of leaders whose judgment on hiring is trusted enough to hold the bar across levels and functions. I earned the designation and served on hiring panels across the organization — not just for my own team, but wherever the standard needed keeping.

When Enterprise Support launched its manager-excellence recognition program, I was first. The award wasn’t applied for — it was given. It reflected the team culture, the development practices, and the leadership model I’d built inside the org over years. First in the program’s history.

My town’s historic preservation committee reviewed city council agendas by hand — downloading PDFs, looking up block-and-lot numbers, cross-referencing zoning maps. I automated it. A pipeline that pulls meeting materials from the city portal, resolves every property reference to a real address, checks it against historic district boundaries, and assembles everything into a clean HTML briefing. The first time the board used it, the response was: “This is great.” Built evenings. Shipped July 2026.