Building open source developer tools for AI-assisted workflows.

Current project: Domscribe — a build-time plugin that maps every DOM element to its exact source location, so AI coding agents like Claude Code, Cursor and Kiro know precisely what to edit without searching. Works with React, Vue, Next.js, and Nuxt across Vite, Webpack, and Turbopack.

→ github.com/patchorbit/domscribe

→ domscribe.com

Subject matter expert for Amplify Hosting as a whole. Deep ownership across Control Plane, Build System, and Data Plane

• Designed cross-service hostname mapping architecture solving critical data consistency risks across service boundaries; authored 2 design documents shaping implementation direction

• Led end-to-end security review for Control Plane architecture spanning 4 sub-components; authored unified security design document establishing platform security posture

• Designed unified webhook architecture with WAF-based rate limiting; eliminated service-level throttling incidents, reduced app creation failures by 20%; led 4 engineers across 2 teams

• Built ECS Fargate Build System POC in Rust; validated 50% build time reduction and 20% cost savings potential using Step Functions, Lambda, and warm pool architecture

• Led security incident response for Git webhook secrets vulnerability; designed KMS envelope encryption migration with zero customer impact; established dual-key rotation mechanism

• Designed shadow traffic capability in Traffic Gateway; reduced production regression bugs by 50% through safe traffic mirroring;

• Mentored engineers across all 3 Hosting teams; conducted operational excellence training across all of Amplify (30+ engineers)

Joined AWS Amplify Hosting as the team scaled from 6 to 30+ engineers. Became the subject matter expert for the Build System while contributing across all Hosting components.

• Architected SSR Hosting Primitives separating static assets, compute, and image optimization; reduced serverless bundle sizes by 60% and increased SSR framework adoption by 25%

• Designed and launched the Amplify Deployment Specification, the standardized build output format for SSR frameworks; led AWS partnership with Nuxt.js to create their official adapter

• Redesigned deployment upload algorithm from batch to cascading parallel architecture; reduced P99 deployment times by 50%, up to 70% for large applications

• Led Amazon Linux 2023 migration for build infrastructure across 3 engineers; maintained backward compatibility for 100K+ customer builds; achieved 70% adoption within 6 months

• Built comprehensive observability system (CloudWatch metrics, X-Ray tracing, alarms); improved build system reliability by 50%; surfaced as org-wide operational excellence win

• Contributed code to every Hosting component: Control Plane, Build System, Traffic Gateway and Compute Service

• Promoted to L6 (Senior) in 2.5 years; fast-track promotion reflecting technical leadership and cross-team impact

I was the lead architect at Whiteboard, a mortgage CRM software company that caters to the mortgage industry. My main responsibility was leading the development of a distributed, scalable and highly-available cloud-based web application.

My role here involved designing full stack applications, implementing git workflows, establishing development standards, leading architecture discussions with the rest of the team, coordinating with QA engineers and release management.

A key project I delivered involved an architectural overhaul of our backend application by re-writing services in NodeJS. Using Loopback, Typescript and implementing Dependency injection as the primary design pattern, allowed me to establish a standard of coding that was easy to unit test.

I played the lead role in the complete infrastructure overhaul of the application from dedicated hosting to a containerized cloud environment. This involved breaking the application into smaller services and separating the concerns of each piece. I setup a Docker swarm cluster, with each of the smaller pieces deployed in their own isolated containers while talking to each other through a message broker like RabbitMQ. Combining this with AWS CloudFormation, our entire infrastructure is versioned and tracked as simple text files. This allowed me to setup isolated testing environments with the concern separation at the infrastructure level.

I also played the lead role in setting up our continuous delivery pipelines. The pipelines allow for automated deployments, with each change that's tracked by Git triggering a new build. Each build goes through our testing environments before going into production. I also architected the automation of unit and end-to-tests which are integral to each of our pipelines.

I played an integral role in designing services that enable third-party integrations with mortgage data providers such as Calyx and Ellie Mae Encompass using NodeJS, Loopback and Typescript.