Impact at a Glance
Day One: January 7, 2017
I started at One Medical on January 7, 2017. By the end of that first year, I'd tackled one of the biggest technical challenges in the codebase: migrating over 50,000 lines from CoffeeScript to ES6.
It was methodical work—converting resources, directives, and controllers one by one while keeping the production system running. This was healthcare software. Downtime wasn't an option.
Four Phases
2017-2018: Learning the Domain
I spent my first two years deep in the clinical charting system. CoffeeScript migrations, prescribing workflows, 2FA integration. I built the foundation—both technical and domain knowledge—that everything else would build on.
- Built push-enabled 2FA for prescribing
- Migrated 17+ modules from CoffeeScript to ES6
- Started AngularJS → Angular hybrid rewrite
- Established build tooling and linting standards
2019-2020: Building Shared Infrastructure
I started seeing patterns across our React applications. Instead of solving the same problems repeatedly, I extracted them into shared libraries. In 2020, I also removed 32,000 lines of code—sometimes the best work is deletion.
- Created developer tooling library
- Created design token system
- Built insurance verification workflows
- Major refactoring: -32K lines removed
2021-2023: Platform Engineering
I shifted from building features to building systems. Created a full design system. Became technical lead for the scheduling platform. Started thinking about how to enable teams, not just ship code.
- Created component library and design system
- Became technical lead for scheduling
- Deepened backend skills (Ruby/Rails)
- Built internal admin tooling
2024-2026: Technical Leadership
I led organization-wide migrations. GraphQL across 10+ applications. Angular modernization. AI treatment recommendations integrated into clinical workflows. Built tooling that prevents bugs before they reach code review.
- Led GraphQL migration across organization
- Integrated AI treatment recommendations
- Modernized Angular (signals, control flow)
- Built automated code review platform
Libraries & Shared Infrastructure
I created and maintained multiple shared libraries used across One Medical's engineering organization:
Developer Tooling Monorepo
Monorepo of shared tools for React applications. I created and contributed to multiple packages:
- react-style-reviewer: Automated code quality enforcement with GraphQL, CSS, and TypeScript analyzers
- react-rum: Real User Monitoring integration with CloudWatch
- auth: Authentication hooks and utilities for role-based access control
- profile: Hook for fetching internal user profile information
- graphql-auth: GraphQL authentication support
- graphql-subscription-hook: GraphQL subscriptions for real-time updates
- elasticsearch-hook: Elasticsearch query utilities
- micro-frontend: Micro-frontend architecture tools
- eslint-config-react: Shared ESLint configuration
- prettier-config-react: Shared Prettier configuration
- deploy: Deployment scripts and utilities
- react-hooks: Basic hooks and helpers
Design Tokens
CSS custom properties for theming. Single source of truth for colors, spacing, shadows. Supports dark/light modes with runtime switching.
Component Library
Shared React components, icon library, patient entity components. Used across 5+ applications for consistent UX.
GraphQL Schemas
Shared GraphQL schema definitions used across frontend applications. Enabled type-safe API integration.
CircleCI Orbs
CI/CD automation for React applications. Standardized build and deployment pipelines.
Migrations
I led or contributed to multiple major technology migrations while keeping production systems running:
4,503 Code Reviews
For every pull request I authored, I reviewed three others. That's 4,503 code reviews over nine years—about 500 per year, or roughly 10 per week.
In 2019, I reviewed 672 PRs while authoring 171. That 3:1 ratio stayed consistent throughout my time there.
AI Integration
My most recent work focused on integrating AI into clinical workflows in two major ways:
AI Treatment Recommendations
Integrated ML-powered treatment recommendations directly into the clinical charting system. This meant solving problems like:
- Medication deduplication using composite keys
- Soft stop vs hard stop logic for clinical safety
- Patient-reported data integration
- Feedback loops for model improvement
Embeddable AI Assistant Widget
Built an embeddable AI chatbot widget for the clinical charting application. Technical challenges included:
- Shadow DOM architecture for complete style isolation
- Real-time streaming with Server-Sent Events
- Context management across host app and widget
- CSP-compliant styling with runtime CSS generation
62 commits, 22.5K lines. Production-ready widget that can embed in any web application.
Healthcare software is different. Patient safety is paramount. Every decision has to account for that.
Major Projects
Clinical Charting System
The primary clinical tool used by all One Medical clinicians. I worked on prescribing workflows, insurance verification, treatment recommendations, and the AngularJS to Angular migration. Led the Angular modernization effort with signals and control flow syntax.
Scheduling Platform
Built multi-select calendar, real-time event updates, timezone handling, booking rules engine, and office/provider search. Became technical lead in 2023.
Backend Services
Backend for all One Medical applications. Led GraphQL schema design, insurance carrier APIs, contact suggestions, and treatment recommendation backend.
Internal Admin Tools
Built admin interfaces for operations teams. GraphQL migrations and internal tooling.
Impact Beyond Code
Writing code is one thing. Multiplying your impact is another:
The libraries alone probably saved hundreds of hours of duplicate work. The code review tooling prevents bugs before they reach production. The migrations enabled teams to move faster.
What I Learned
Nine years. 1,518 pull requests. 250,000 lines of code. But the numbers don't tell the whole story.
I learned that the best code is often the code you delete. That migrations matter more than greenfield projects. That building tools for your team multiplies your impact more than any individual feature.
I learned healthcare software the hard way—by maintaining it, refactoring it, and keeping it running while thousands of patients depended on it.
This is my work at One Medical so far. It's a fraction of my career, but it taught me what matters: building systems that last, enabling teams to move faster, and never forgetting that real people depend on this code working.