RanchML is a GitOps-powered ML model deployment platform that transforms the way companies ship models to production. The platform enables teams to deploy models in minutes rather than weeks while maintaining enterprise-grade reliability and security. (Go, Kubernetes, React, TypeScript) • Designed and building a cloud-native ML deployment platform using modern infrastructure tools • Developing custom Kubernetes operators for ML model lifecycle management • Implementing observability stack with OpenTelemetry, Prometheus, and Grafana • Creating GitOps workflows with ArgoCD for reliable model deployment and rollback • Designing scalable, secure infrastructure leveraging cloud-native technologies • Building developer platform with Backstage for documentation and service discovery

I inherited a team with communication, collaboration and morale issues, led them through a round of downsizing and coached them into a high-performing, tight-knit team with outstanding morale. (ECS, Kinesis, Postgres, Python, Rails, React, TypeScript) • Scaled backend & infrastructure to process the massive number of incoming AI-generated resources • Delivered fraud detection, partner integration API, payment automation & performance optimizations • Led frontend and QA through CS50 and other training following layoffs; all became full-stack engineers • Drove developer process improvement and workflow optimization, improving productivity 20%+ • Implemented observability with OpenTelemetry and Honeycomb, enabling data-driven optimization

I worked primarily with startups struggling to launch, coaching and educating while addressing process and architecture hurdles to ensure thriving humans as well as reliable, scalable and maintainable systems. • A “shopping-monetized TikTok” startup was failing to deliver its promised public beta. The backendhad been built by engineers unfamiliar with working fast, production was plagued with errors anddevelopment would fail intermittently causing the team to believe they’d broken something and wastecumulative hours troubleshooting. Half-hour-plus deploy/rollback times had created a culture of fear.Duplicate effort and work on stale projects were both major issues. (Python, Rails, React, Rust) -– Implemented planning and review meetings, eliminating duplicate effort and stale tickets -– Rewrote backend, reducing codebase 70% and resolving 100% of ongoing errors -– Drove successful effort to increase test coverage 500%+ & reduce deploy/rollback times 80%+ -– Designed activity feed algorithms and built feed backend -– Led team to successful public beta launch • Partnered with ski resorts & sports retailers on a form automation & loyalty app (Go, React) • Delivered a long-delayed DBaaS product for a mobile database startup (Kafka, Scala) • Developed an online ordering and logistics backend for a restaurant chain startup (Go, Google Cloud)

I was a non-founder executive responsible for everything from communications to design to engineering to managing our relationships with Apple and investors. Apple named Glass Planner a Best New App, anEditors’ Choice and featured it several times with large banners in the App Store’s Productivity section. (Objective-C, Scala, Swift) • Directed all aspects of the company through explosive growth while leading backend development • Built syncing system to attach arbitrary data to third-party calendar objects • Partnered with App Store editors to select and implement Apple showcase features prior to iOS releases • Drove implementation of continuous delivery, increasing backend release frequency 90% • Led automation of iOS release processes with Fastlane & custom tooling, dropping app release time 80%

I led initiatives to address scaling issues with legacy systems processing financial transactions for hundreds of banks and credit unions, driving the development of streaming data pipelines. (Kafka, Rails, Scala) • Developed novel key strategies to resolve scaling issues, reducing DynamoDB errors 90% & costs 20% • Uncovered and resolved time zone defects in 20% of test suite, sped up test suite runs 400%+ • Partnered with Typesafe/Lightbend on migration from Rails to Scala and new data processing pipeline • Wrote moe library for working with DynamoDB at scale: compound keys, mirrors, time-series tables, etc

Abstract Ranch developed internal SaaS products and provided infrastructure and web application development consulting services. • Designed infrastructure management system for Autonomy and coached relevant teams • Built microservice management tools for Fidelity Life, decoupling services & reducing errors 99%+ • Increased shopping cart speed by 80% and decreased CMS synchronization times 99% for LandCentral

I built product licensing and wholesale catalog management apps for major brands. During my tenure I also led the migration of infrastructure from manual on-site management to a container-based hybridinfrastructure-as-code model. • Managed products for Boise Cascade, Brand Sense Partners, CBS, Fox, Jessica Simpson and MGM • Designed pre-Docker/Kubernetes/Terraform container orchestration system, reducing infra costs 85% • Wrote new algorithm for decision tree DSL, improving performance 200%+ • Implemented Rails engines and extracted shared code into Ruby gems to mitigate duplicate effort • Rewrote CSV processing, improving performance 65% and resolving issues uploading very large files

FieldReport was giving away the world’s largest-ever cash prize for a single piece of writing. We were hailed as “The American Idol of Blogging” by the New York Times, National Public Radio and Time Magazine. Using Ruby on Rails we leveraged EngineYard, GitHub, Lighthouse and Campfire to facilitate collaboration and eliminate the need to hire operations staff. • Led development of the FieldReport web application • As editor-in-chief worked closely with major publishers & agencies like Condé Nast & William Morris • Grew engineering team from three to ten, implemented agile processes and remote-first work policies • Reduced infrastructure costs 90%+ migrating to the cloud and leveraging SaaS-based IT products

My team designed and managed the multi-petabyte, globally-distributed storage infrastructure of a major media app. I initiated and led development of a next-generation monitoring and automated management infrastructure; the storage organization was subsequently acquired by Savvis. • Built a web-fronted management, monitoring & self-healing system that eliminated 80% of manual effort • Identified and exploited undocumented APIs to enable Linux management of Isilon clusters

I was the manager and senior technical member of LDC’s systems group. I drove the transition of all computing and network services, formerly outsourced to the School of Engineering, to an internally-sourced model in order to support petabyte-scale data collection and ML model training. • Built a pre-AWS self-service storage & compute cloud, reducing costs 70% and provisioning time 99% • Donated surplus cloud capacity to other ML research organizations around the world • Worked with ML team to build a Hadoop-like distributed filesystem and processing platform • Wrote backup software 300%+ faster than any commercial product, wrote tape library drivers • Implemented data center automation and 100% OOB management, enabling team to work remote • Ran >12T volumes in 2T limit days, resolving OS, file system & driver issues at bleeding edge of scale • Built a custom filesystem view generator to support legacy ML tools with hardcoded paths • Developed Hammer configuration management system, decreasing deployment times 99%+ • Built novel hard disk duplication system, upping capacity&throughput 800%+ over commercial solutions