Skip to main content

Portfolio

Proof from the code most engineers inherit and dread — legacy pipelines and brittle services made correct, documented, and safe to build on.

The work spans two distinct problem spaces: healthcare technology at Oracle Cerner, where correctness and patient data integrity are non-negotiable, and cloud FinOps engineering at VMware's CloudHealth, where cost reporting accuracy directly affects customer business decisions. Both required the same discipline — write systems that work predictably and that other engineers can maintain.

Career timeline

5 roles

Impact case studies

Azure VM Metrics Pipeline: CloudHealth Modernization

Drove modernization of an outdated Azure VM metrics pipeline at VMware's CloudHealth by implementing the latest Azure Monitor API — restoring 100% data accuracy and enabling cost-efficient decision-making for thousands of customers.

Business impact: Thousands of CloudHealth customers regained accurate Azure VM cost and utilization data, enabling informed FinOps decisions after a period where the outdated API integration had introduced data gaps.

Azure Monitor APIRuby on RailsJavaMulti-cloud platformCloudHealth pipeline

Key decisions

  • Identified the root cause as the outdated API version rather than a downstream data processing issue — directed the fix at the interface
  • Validated 100% data accuracy across customer accounts before marking the initiative complete
  • Documented the Azure Monitor API integration fully so the next update cycle wouldn't require the same investigation
  • Coordinated with the CloudHealth platform team to ensure the upgrade didn't affect adjacent Azure integrations
  • Measured customer impact quantitatively — thousands of accounts regaining accurate FinOps decision data

100%

Azure VM cost data accuracy restored

Thousands

customer accounts impacted by the modernization

Healthcare Incentive ETL Pipeline at Oracle Cerner

Engineered a data processing pipeline using Apache Crunch, Sqoop, and Oozie in Java to power Oracle Cerner's incentive-based healthcare workflow — ensuring patients receive incentives for completing annual physical tests.

Business impact: Replaced a fragile workflow with a robust, tested ETL pipeline that reliably triggers incentive delivery, directly affecting patient engagement and healthcare compliance outcomes.

Apache CrunchSqoopOozieJavaHDFSMySQLRuby on RailsDelayed JobOpenAPI/SwaggerGrapeApache MesosSpinnaker

Key decisions

  • Chose Apache Crunch + Sqoop + Oozie as the ETL stack for reliability under the data volumes and scheduling requirements
  • Built the Questionnaire service with Two-Phase Commit (2PC) to ensure distributed consistency across dependent services
  • Used Delayed Job for notification delivery to avoid blocking the main request cycle — appointment and lab result notifications are time-sensitive but not synchronous
  • Documented and exposed all APIs through OpenAPI specification via Grape — so consuming teams had a reliable contract
  • Deployed using Mesos and Spinnaker for repeatable, consistent production delivery

6yr

production Rails and Java engineering at Oracle Cerner

2PC

Two-Phase Commit for distributed service consistency

Ruby Codebase Quality and Azure FinOps Reporting

Led systematic Ruby code refactoring at VMware's CloudHealth to reduce technical debt and align the codebase with Rails standards — alongside building Azure FinOps data grouping in Java and Scala to improve Cost History Report performance.

Business impact: Meaningful reduction in team technical debt and improved codebase navigability for incoming engineers, alongside measurable latency and customer support ticket reduction on Azure cost reporting.

Ruby on RailsJavaScalaRabbitMQApache KafkaAWS S3CloudHealth

Key decisions

  • Approached refactoring systematically — identifying the highest-leverage areas of debt first, not refactoring at random
  • Built the Azure data grouping logic in Java and Scala rather than extending the Ruby service — keeping concerns separated and each service maintainable
  • Built the S3 Observability Tool to detect file corruptions proactively — catching billing cycle issues before they reached customer impact
  • Treated every refactoring PR as a documentation opportunity — ensuring refactored code was readable without requiring inline comments
  • Tracked latency and customer support ticket volume as outcome metrics for the Cost History Report improvements

Reduced

latency and support tickets on Azure cost history reporting

Systematic

Rails debt reduction — not a one-off rewrite

You've seen the work.

Next: How I think