Lead eng design and development of an LLM-based Meal Planning feature, from initial prototype at a hackathon through full productionalization in less than 3 months. Testing in an A/B test showed a 4% lift in our team's core OKR, and an overall 2% lift in DAU in the test group.

Approached iteration on the feature with a funnel analysis with the key user engagement steps, and was able to lift the first step by 10% with targeted engineering improvements.

This effort introduced a pattern of working with the OpenAI Assistants API, and democratized AI development by establishing internal best practices for working with LLMs in our normal SDLC.

Advised on engineering designs for several other projects, emphasizing rapid prototyping an iteration and minimizing time-to-value. Developed several new microservices on our best-practice toolset, including GraphQL, protobuf based grpc interfaces, and ent-based generated database schema code.

This role introduced me to mobile development and member-facing features; as starter tasks I identified and fixed several latency issues on our homescreen (>90% performance improvement), and brought in PDF viewing to enable labs in app.

Fullstack role on the team responsible for our internal EHR system. Contributed to and led rollout of major effort to recreate our clinical intake workflows in the internal system vs an external vendor that was not well suited to the purpose. Effort was delivered ahead of deadline, with major coordination with product and clinical stakeholders for UAT testing and bug fixing.

Tech lead project to redefine "baseline" lab values in our system, working cross-functionally with partners on the analytics team, as well as our external vendor for lab data. This allowed deprecation of several data pipelines and ensured all analytical and operational use cases were using the same data.

Worked heavily with FHIR objects and presented internally to engineering to help educate on the purpose of FHIR and what sorts of data are appropriate to store in the FHIR store.

Part of a large effort to migrate our entire system onto Google Cloud Platform. This gave us greater control over the details of our systems' configuration, and unlocked functionality via Google's various managed cloud services. As part of this migration I:

Post migration, deprecated several old tools made redundant by GCP, improving cost and security. Wrote a healthchecking proxy service in go, enabling use of external healthchecks from Google Cloud Monitoring. Put together and documented the process for bringing up new services, including a basic sample repo, and presented that information. This has since allowed 3 additional new microservices to be created.

Team reorged into SRE, where I lead load testing efforts, unlocking stability for a big period of growth, and was the internal export of our observability platform Honeycomb.

Developing Kubernetes and CloudFoundry BOSH based PaaS products for internal and external customers. Automated deployment of all major system components using Jenkins, scripts, and AWS SSM Agent. Deployed Prometheus as metric-level monitoring solution, configuring custom dashboards, alerts, and certificates using custom BOSH release

Sole developer building out MVP1 of an internal monitoring product. Backend coded in go with a microservices architecture, frontend in javascript with React. Responsible for gathering and prioritizing requirements from various managers interested in this project, building out features to add additional data sources and insight management, making architectural and UX decisions, and regularly presenting and soliciting feedback from stakeholders.