At Cloudflare · Building Endium · Previously Google
Here’s what I’ve been working on.
Customer build · Cloudflare
I worked with a healthcare team on an AI-guardrails MVP. We got together to work through the design and build the software.
These are our working notes: how the checks fit together, what each one returns, and how to tell whether something passed.
Working on the MVP together.
A check ran. But did the content pass?
Figuring out which checks run, and when.
Solutions engineering teams · In progress
An engineer may need months of account history to answer one question.
I’m building an engineer agent at Endium that brings that history and the product sources into the reply workflow. I built the account memory, Slack integration, draft editor, and approval flow.
Built so far: A question in Slack becomes a draft with sources attached, ready for an engineer to review, edit, and approve.
Inside the build
Cloudflare sales engineers · Early internal app
The team’s account history was scattered across notes, files, and conversations.
I built Account Memory Agent for Cloudflare’s sales engineers to bring that history together. I connected transcript processing, semantic search, meeting prep, and a daily briefing.
The result: One place to find earlier decisions, prepare for a call, and see who still owes a follow-up, including when another engineer takes over the account.
How I built itEndium · AI front desk in progress
Someone reaches out about a project. The team needs to understand the request and arrange the next conversation.
I’m building Front Office Orchestrator for Endium first. I built the conversation tools, booking checks, calendar integration, and handoff so a person can step in with the details already collected.
Built so far: A visitor can explain what they need, choose an available time, and get a confirmation only after the calendar booking succeeds.
Inside the build
We launched Morph.Systems at TechCrunch: an AI platform whose agents turned an app idea into product requirements and then built the application.
I like a problem I haven’t solved before. Especially when solving it makes someone’s day easier.
That could mean spending less time on repetitive work, making sense of something complicated, or trying an idea that used to feel out of reach. I enjoy working through those possibilities with people, then making something they can use.
That’s the work I do with customers at Cloudflare and through Endium, the company I started. I’m involved from the first conversation through the prototype, the implementation, and the things that need fixing along the way.
I’ve spent more than 20 years in engineering. At Synergent, I designed and implemented two datacenters supporting more than 50 credit unions. At Google, I worked on the GPU infrastructure, networking, and storage behind AI training and inference. Today, I’m putting that experience into applications and agents people can use.
I help customers build on Cloudflare and work with our product and engineering teams when something gets in the way. Recent work includes an AI-guardrails MVP with a healthcare team. To unblock container-platform testing, I shipped three PRs and helped get six issues fixed and merged in eight days.
I also designed and led the first day of technical training at our 2026 technical summit, spoke to 300 field engineers, and trained more than 200 engineers on the AI stack.
I started Endium to build AI tools, automation, and custom software for businesses. I work with clients to decide what to build, then handle the implementation and delivery. I’m also building an engineer agent that remembers customer context and prepares replies for review.
Visit EndiumI designed GPU and HPC infrastructure for training and inference, built prototypes, and led architecture reviews from design through deployment. The work covered compute, networking, storage, security, Kubernetes, and Slurm.
I contributed to more than 60 new customer wins in one year and helped an inference provider grow consumption by more than 300%, adding $500K in revenue.
I helped host the Techstars Boston Fall 2024 Demo Day at Google’s Cambridge office, where startups pitched their companies.
I helped customers plan their move to the cloud, design their infrastructure and data systems, and test ideas with working prototypes. The plans had to fit how their teams actually operated.
I worked with enterprise teams on network security, including zero-trust networking and microsegmentation. I helped technical leaders and executives understand the options and plan the implementation.
I designed and implemented two full-stack datacenter environments supporting core technology services for more than 50 credit unions, across compute, storage, networks, security, and operations.
On the founding team at Morph.Systems, I helped launch an AI app-building platform at TechCrunch. Its agents turned a plain-English app description into product requirements and then built the application. I’ve also worked on voice agents and matching systems with Aquarix AI.
If you’re working on something interesting, want to collaborate, or just want to say hello, drop me a note.