Howdy
I build software and infrastructure across cloud computing, local AI, and privacy-focused tooling.
I'm drawn to open source and local AI because they give you more control and fewer dependencies: running models on your own GPUs and NPUs, owning your authentication and communication, and not being locked into a single provider. I like taking complex systems and making them easier to deploy, operate, and use.
I work at Microsoft, writing code for customers in private repos. On my own time, I'm exploring:
- Local and on-device AI across CPUs, GPUs, and NPUs (oddo-desktop)
- Generative AI runtimes and model optimization (Lemonade)
- Cross-platform desktop applications (Kava)
- OIDC/OAuth2 and Passkey authentication gateways (Vouchrs)
- Rust, C#, C++, Python, and TypeScript (across these projects)
- Lemonade - Local AI server and embeddable runtime for running models on GPUs and NPUs
- signal-call-tunnel - Private, encrypted calling infrastructure
- kmscon - Kernel-based virtual terminal and console emulator
- Linux kernel - Bluetooth btusb driver support for MediaTek MT7925 hardware
I spent a couple of decades getting here the long way.
I started out in the late 90s at an ISP owned by a telephone company, then spent years as a systems administrator across very different worlds: a Vermont catalog retailer, oilfield manufacturing at Halliburton, pharma at Genzyme, and federal security compliance at MIT Lincoln Laboratory. In between, I deployed and ran VSAT satellite communications networks in southern Afghanistan.
In 2015 I joined Xamarin just before the Microsoft acquisition. From there I moved from operations into engineering: multi-cloud automation and SaaS product dev at RedPoint Global, cloud architecture leadership at IBM, and cloud-native consulting at Microsoft, where I built container and Kubernetes IP and modernized applications across finance, retail, manufacturing, and healthcare.
Along the way I architected large-scale Azure and Kubernetes platforms for autonomous vehicles at General Motors, and eventually returned to Microsoft as a Principal Software Engineer, where I now design and build agentic AI systems for Fortune 500 customers.
I write about Azure, GPUs, containers, software development, and the lessons that come from building things: