Enjoyed working with the team at Solo!Got to help blitz a bunch of SOC2 compliance automation, GCP Kubernetes gitops, gitlab CD, and Javascript app dev features.Still has that cool small company feel, and I found the product features really innovative.

It was an absolute pleasure to build with the team at MX. Besides a relentless focus on the human-centric goal to make people financially strong, I found their software architecture to be par excellence.However, with thousands of VMs across two datacenters, where provisioning an instance took on the order of an hour, the infrastructure and deployment systems, while well architected, were showing their age and hampered the team’s ability to move quickly (we quoted dev a month to bring new VM services online).I was tapped to lead the enormous task of modernizing our infra and deployent system, and within a year we had 7 kubernetes clusters between two bare-metal DCs and GKE, with nearly 5000 cores and 12TB of RAM. I then additionally took lead of and mentored a team of 7 engineers focused on porting applications to containers, and getting them into our clusters.Ultimately we were able to reduce rolling application updates from hours to minutes with robust failure and rollback modes, and the onboarding of new applications was reduced from weeks to days or hours. Not only did this shift allow developers to iterate more quickly, it significantly improved developer experience by removing massive amounts of cognitive load from the whole SDLC.Some key aspects of the design and implementation:- Kustomize-based gitops repo rendered into clusters with ArgoCD- Custom C++ istio filters for offloading TLS origination for postgres and NATS- Custom k8s operator to sync healthy VM service instances to istio from Consul- Custom Argo render pipeline to decrypt inline secrets in-cluster- MetalLB for westbound connectivity with zone delegation to kubedns from core bind DNS servers- Autoscaling, including scale-to-zero based on meaningful metrics like message queue size- Hyperconverged rook.io managed ceph storage with S3 emulation, snapshots, and cross-cluster replication

For nearly a decade, I was responsible for all technical aspects of bootstrapping the world’s first wireless, mobile-first, cloud-based physical access control system, with a goal of bringing modern technology and SDLC to bear on the problem. I oversaw the design and manufacturing of our in-house hardware and software, which involved managing a team of over 20 engineers.On top of a continuously delivered SDLC and radical developer ownership, we adopted Kubernetes early-on for fault tolerance, instant zero-touch deployment of tested code, and observability across services at all levels. This auto-scaling system was intentionally built to be provider agnostic to allow for optimizing performance, availability, and cost.An IoT runtime and delivery platform is core to the system, centered on a custom embedded Linux container-OS and purpose-built hardware--we call a cloud node--which critically maintains functionality in the face of internet connectivity issues. Uniquely, we manage this fleet like so many servers, using Ansible rather than complex multi-partition schemes. Today this hosts ~3x10K nodes.A custom delta update system for container images is crucial to CD over our worldwide network of devices on unknown and often difficult network toplologies, with patches on the order of 15KB vs 100MB+ images.As leafs to the cloud node, we manufactured intentionally simple interface boards to provide locality and focus on the physical aspect of access to doors, gates, elevators, and etc.An OIDC IdP uses claims-based tokens for flexible authentication schemes like federation and controlling hardware locally without the internet. We were the first with any 2-factor authentication and remain the only to support hardware tokens.I take immense pride in the role I took in building this company, and am grateful for the amazing team supporting us throughout the journey.The most recent transaction valued us at ~$30M, with SaaS multipes now estimating north of $100M.

With a love of all things IT I accepted a position at Ferrari Color to design and implement its IS strategy. Much like the early years at 1-800 Contacts, Ferrari Color is a progressive company that sees technology as a point of differentiation in its market segment, and needed someone that could build on that core tenet.As Ferrari Color had seen competitors fail during "The Great Recession" we fought to be consistent and provide a level of service to customers that could not be matched by competitors.As part of the IS strategy, I worked to modernize the network and management systems, implement server consolidation through virtualization and build server infrastructure to handle the new systems and processes.I played a key role in selecting, implementing, and deploying a new enterprise management system to handle everything from manufacturing process, and CRM, to finance. Once implemented this system also required a number of customizations and extensions to meet unique business needs.As a custom product manufacturer we produced, deployed, and managed nearly a dozen B2B web sites for fulfillment of customer product produced in-house. These sites required a number of back-end systems for ordering and distributing those goods to the parters' locations.The B2B intellectual property allowed Ferrari Color to differentiate itself in the market and be very responsive to customer needs, giving the customer agility to excel in their field and provide a premium experience to their customers.In addition to the B2B sites, Ferrari Color also provided a number of direct-to-consumer sites on behalf of partners. A couple of public examples are Time Cover Store and SI Covers.My years with Ferrari Color were transformative; with a primary role in these initiatives I had a direct hand in enabling capacity for growth to expand manyfold.

1-800 Contacts was ahead of its time. Led by Jonathan Coon, aggressive use of technology was a primary strategy used to streamline its processes and provide unrivaled service to its customers. This resulted in unrivaled loyalty and industry partnerships that drove very rapid growth, making them the largest direct-to-consumer contact lens distributor in the world; larger than all competitors combined.In mid-1999 I joined one other developer to design and program the company's first e-commerce site. By the end of 2000 the site accounted for nearly 40% or $54M of the company's gross revenue. The following year I was tasked to work with the marketing team to revamp and completely rewrite the site. By the end of 2002 the site was processing $71M in sales.At this time 5 developers, a 3 person desktop support team, and a single network operations person made up the whole of the IS department.I created the initial version of a computerized FAX system to automate eye-doctor prescription verification, this system's genesis ended up being absolutely crucial to 1-800 Contacts' ability to sell contact lenses as we worked to revamp prescription laws to benefit the consumer.Because of our small IT team I also personally initiated a number of systems refreshes including migrating our directory services from Novell to Active Directory, and moving our messaging infrastructure from Novell Groupwise to Microsoft Exchange.By the time I departed, we had grown the IS department to consume an entire floor of a second building. That department included a multi-million dollar in-house datacenter, a Business Intelligence and Data Warehousing team, QA team, Infrastructure team, Telephony Systems team, Business Software team, and Web Software team.By then more than half of the company's $500M revenue was flowing through its e-commerce systems.

Computers, BBSs, and programming have always interested me, from reading books on DOS to helping install and configure the first network in our school district. After getting a small taste of the power of interconnectivity, what I really wanted to get into was this new "internet" thing that I had started hearing about. As fate would have it, a Boeing engineer used his life-savings to start an ISP in my little home town.As a high-school sophomore, I went down to his office and offered to work in exchange for learning anything and everything I could. He eventually hired me, and we grew the service to a double-digit percentage of the town's population.I had no exposure to Unix at this point and had to figure out how to do eCommerce using CGI and Visual Basic on a Netscape webserver that we had running on Windows NT 3.51 (what's an environment variable??)Within a few months we had created an automated onboarding system that rivaled the AOL disk experience that were flooding people's mailboxes at the time. The user could grab our disk, put it in their computer and it would autorun an installer to install a web browser and TCP/IP stack, configure a connection, and dial into our walled garden to access the signup flow.At the conclusion of the signup process the system allocated an account in our back-end billing system, set-up recurring billing, and transparently configured the dial-up credentials on the user's computer to immediately access the internet.We started hosting and I designed and taught classes for customers and residents of our town on how to use the Internet, email, and create websites. From that exposure I grew a small consulting business assisting people and businesses with computer hardware and software problems at their homes and offices.The things that I learned in the trenches about computers, hardware, networking, and user-interaction/pain provided an incalculably valuable perspective for my eventual career choice as a software engineer.