About

I’ve spent most of my career working on systems that operate under real-world constraints, where data is incomplete, systems do not line up cleanly, and things do not always go as planned.

My work has spanned consulting, building products, and operating infrastructure at scale. Across all of it, the common thread has been making complex systems understandable and workable, especially under pressure.

Early Work and Consulting

Before joining Google, I ran a consulting practice under Innovative Data, working with companies across manufacturing, logistics, and operations.

Some of that work included:

extending CAD systems to embed manufacturing metadata for scheduling and optimization
translating natural language descriptions into structured production processes
building predictive models from sales data to anticipate customer demand
designing systems to improve logistics efficiency and route utilization
contributing to control systems for robotic testing processes

PubTV

I also built PubTV, an online analytics and reporting platform combining Nielsen peoplemeter data with Tribune Media scheduling information for public television.

The system supported reporting, direct querying, and custom research, and was used by organizations across public television to understand audience behavior and programming performance.

PubTV was later acquired by Trac Media.

Writely

Around the time of PubTV’s acquisition, I worked with Writely, where I designed and implemented multi-location, redundant, and secure infrastructure to support early growth.

This included helping the system handle large traffic spikes and supporting the transition to Google infrastructure after the company was acquired.

Google

At Google, I worked on large-scale infrastructure systems, often in situations where things were complex, unclear, or actively going wrong.

Over time, I became someone teams would pull in during high severity incidents. Not because I had all the answers, but because I was good at helping people slow things down, understand what was actually happening, and move toward a solution without making things worse.

Some of the systems I worked on included:

Last mile machine delivery system: processed millions of machines and configured the definitions that determined their roles inside Google’s datacenters.
Storage upgrade automation: increased storage upgrade capacity from hundreds of petabytes to tens of zettabytes per year.
Abnormality detection and uptime measurement: helped measure reliability across very different products and was presented quarterly across Google.

I was not usually the loudest voice in the room, but I was often the one keeping things steady.

Patent

I also worked on a side project exploring how to identify and separate overlapping conversations using audio-only input.

This led to a patent: “Turn-Taking Patterns for Conversation Identification” (US20140081637A1).

The work focused on inferring how simultaneous conversations map to individual participants based on conversational dynamics. It was later cited by patents from multiple organizations working on conversational analysis and communication systems.

Home