Hi, I’m Ruth :)
Senior Fullstack Engineer (8+ YOE) specializing in high performance React, TypeScript, Node, & Python
Co-founded Reactime, a state debugging devtool with 2,200+ GitHub stars, nominated for a React Open Source Award.
Most recently at Tomorrow Health, helped build an agentic system to chase documentation and used AI to automate complex workflows to reduce operations time. Previously architected React and Python / Node services for a greenfield supply chain platform at Premier Inc. during a $2.6B acquisition. At Medidata, built metric benchmarking and ML-projected visualizations for an AI powered clinical trial platform serving 11M+ patients, notably helping Moderna bring their COVID vaccine trials faster to market.
Things I have built
Drydock
Drydock
An agent that migrates code, where the test suite decides
Point it at a mechanical migration and it works one file at a time: propose a replacement, run that file’s tests, keep the change if they pass, feed the failure back if they do not, then revert and flag the file for a human. The agent never reports its own success. Writes and commands go through a policy engine with separate read and write allowlists and a budget on steps, tokens, dollars and wall clock.
✨ The demo page replays a real run and lets you edit any file and run its actual tests in your browser
Snowline
A wallet-side project on Avalanche
In progress. Core wallet backend work on Avalanche, written up here once the first milestone is shippable.
Thyroid Reproductive Hormone Health Explorer
Thyroid Reproductive Hormone Health Explorer
Visualizing what the research says about hypothyroidism and women's health
An interactive walk through the cascade from thyroid dysfunction to reproductive effects, where each node opens into the evidence behind it.
✨ To be featured on a reproductive fertility doctor's podcast
Reactime
Reactime
Open source React devtool
A Chrome DevTools extension for time travel debugging of React state. Co-founded it in 2019 and it has picked up 2,200+ stars since.
✨ Nominated for a React Open Source Award in 2020
Where I have worked
Tomorrow Health
- Co-planned and co-built the migration of THEA, an autonomous AI voice agent that chases the documents needed to bill medical equipment orders, from a Google Apps Script and Sheets prototype to a production Python platform.
- Reverse engineered a 5,000 line script into its load bearing agent rules and rebuilt them as Celery cron jobs behind a GraphQL API. Piloted with 10 supplier companies over one week each, 75% converting to paid contracts, generating roughly $2,000 in weekly revenue per company.
- Designed and iteratively built the React and TypeScript admin UI for THEA, collapsing 5+ tabs of patient order data into a two column filterable doc chase status and detail view, cutting human escalation review time from 60 to 10 minutes per case.
- Built the safety gates that let THEA operate in a regulated billing environment, gating billing on a single verified document from the doc chase pipeline rather than on call activity, cutting false positive billing errors 88% across pilot companies.
- Built supplier and admin self serve UI for two of six DME routing modernization initiatives, part of a broader effort that cut the routing ticket to order ratio from about 25% to under 20%.
- Built the Gemini classification pipeline and prompt schema behind AI fax intake and ran a phased rollout, driving AI fax adoption to 73% of fax orders and cutting manual offshore completion.
Premier Inc.
- Worked on a greenfield supply chain platform for 4,000+ hospitals through a $2.6B acquisition, mostly on surfacing domain gaps early enough that they did not become release problems.
- Drove architecture decisions on the React/TypeScript frontend and the .NET + C# CQRS backend API contracts, shipping flexible UI foundations ahead of the specs so downstream teams were not blocked.
- Strangled reporting modules out of a legacy .NET monolith into FastAPI services while migrating the frontend to React microfrontends, using feature flags to roll out with no downtime.
- Built a Python automation engine on LangChain and Azure OpenAI (RAG) that bootstrapped local environments and automated git workflows, cutting VM setup time by 88% and saving 480+ engineering hours a year.
Medidata Solutions
- Built the metric benchmarking and ML projection visualizations in D3.js for the clinical trial platform, including the layer used on the Moderna COVID-19 vaccine trials.
- Led the Q2 charting initiative and designed Python service layers that removed repetitive config, shipping two weeks early.
- Refactored HOCs, reducers and data pipelines for maintainability, cutting average cyclomatic complexity by 33%.
- Cut frontend latency by roughly 0.8s by redesigning the API contract and how the client handled the data.
Joy
- Stabilized a rapidly scaling microservice architecture with global error handling middleware in Node, which restored observability and cut unhandled exceptions by 60%.
- Isolated AWS upload failures using telemetry, which made recovery and safe database rollbacks possible during production incidents.
Peacekeepers
- Led a 3 person engineering team to design and launch a platform improving access to legal resources.
- Partnered with the CEO on a greenfield React, Node and Python platform, with FastAPI for data logic, Express for real-time APIs, and GitHub Actions CI/CD built around TDD.
Codesmith EdTech Platform
- Moved core platform infrastructure onto Docker so the team could work remotely through the pandemic.
- Ran public JavaScript workshops for over 100 people, and taught Python and AWS deployment workshops to bootcamp cohorts.
Reactime (Open Source)
- Architected Reactime, a time travel debugging devtool for React state. The D3.js visualization renders state changes by traversing the React fiber tree, with AST parsing to detect hooks inside webpack bundles.
- Published the npm package for state tracking integration with Chrome devtools.
- Built a port based system so you can debug several React applications at once.
Allergy Asthma Sleep Center
- Built a custom patient data platform on AWS with a SQL database, role based authentication and clinical intake forms, replacing handwritten allergy and sleep records.
- Designed the interfaces clinical staff used to enter and retrieve patient information in one place, instead of across separate systems.
- Automated synchronization across previously siloed clinical records, which routed documentation to the right provider and cut hours of manual processing a day.
MBCC (Nonprofit)
- Built and maintained consumer facing web features including dynamic content displays and event pages, across thousands of media assets.
- Built intake and registration pages that integrated with external platforms to generate participant profile pages, improving fundraiser and event onboarding.
- Maintained the databases tracking sponsors, volunteers and participants across multiple annual events.
What I work with
Roughly in the order I reach for them. The last group is honest about where I am still learning.
AI systems
Backend
Frontend
Data & infrastructure
Testing
Picking up right now
Say hello
I am looking for a full time role right now, ideally somewhere the AI systems work is real and the ground is not moving every eight months. Healthcare is where my background is deepest, but any high stakes, regulated domain is interesting to me. If you want to talk, or you just want to ask about THEA, my inbox is open.