01
The internet
1994 to 2005
I've been getting people online since my teens: first a BBS, then an ISP at 17. I rose to VP of Operations at oz.net, one of Seattle's first dial-up and DSL providers, and led its sale to The River.
At The River I architected the integration of the acquired ISPs into one system, then as interim CEO took the company through a turnaround to its sale to a public company.
In 2001 I helped found the Seattle Internet Exchange and was its first President. It made peering free and neutral, and it is now the world's largest free Internet exchange point.
- 30,000+ISP customers at The River
- 371networks on the SIX today
- 3.7 Tbpspeak SIX traffic
02
The cloud
2006 to 2024, and the start of the AI era
I founded Cascadeo in 2006 and was its CTO for nearly all eighteen years, and its CEO when it needed one. I set the direction from day one and ran the P&L for most of its life.
We grew to hundreds of people, almost all engineers, served 100+ organizations, earned a place in Gartner's Magic Quadrant three years running, and won Managed Services Provider of the Year.
We went where the talent was. Most of those jobs were in Manila and rural Legaspi, and building careers there was part of the mission from the start. I sold the company in 2020 and stayed three more years as CTO.
The AI era started here. In 2023 we were among the earliest users of AWS Bedrock, in its pre-release days, and built Cascadeo AI, the first MSP platform to put large language models into live operations and helpdesk ticketing. It led to invitations to speak at AWS re:Invent and the AWS Day One CxO Summit in Singapore, where for many people it was the first working use of LLMs they had seen beyond ChatGPT.
- 18 yearsas founder and CTO
- 100+organizations served
- 3 yearsrunning in Gartner's Magic Quadrant
03
AI
2023 to now
I brought that experience to the University. I am the first Lead AI Architect at UW Information Technology (UWIT), the central IT organization serving all three campuses and UW Medicine, and the primary architect of PURPLE, the University's generative AI platform.
PURPLE went from prototype to campus-wide production in under two years, on a fixed three-year budget, with UW data kept in a tenant the University controls. Its agents are built and refined with a team of staff and student assistants.
The point is to take toil off people's plates, not to replace anyone.
- 100,000+people with access
- 1,200+community-built agents
- 300 to 1,500AI Community of Practice members