core platform and AI in commodities

• Working on AI initiatives at Oscar, primarily focused on improving precision and recall of our risk adjustment pipeline.

• Wrote a new ML onboarding environment from scratch to allow engineers and non-engineers to spin up new ML experiments in minutes.

• Lead on Doscar project, a fully machine-agnostic standardized development environment. With the help of my fellow leads, took the project from 0 to nearly 100% engineering-wide adoption. Specific highlights include implementing Redis and Postgres sidecars and developing a first of its kind PyCharm on x86 Docker solution.

Technologies Used: Docker

• Wrote tooling to assist with migration from Apache Aurora to Kubernetes. Productionized a migration wizard (hosted on Kubernetes) to streamline and partially automate Kubernetes migrations. This tool was used to migrate over 60% of the Kubernetes apps in production today.

Technologies used: Kubernetes, ArgoCD, Helm, Backstage, Bazel

• Designed and wrote the alerting suite for Kubernetes jobs. Spearheaded a migration to Jsonnet alerting which allows for automatic alerting updates.

Technologies used: Prometheus, Jsonnet, Grafana

• Other notable work includes designing and implementing a high availability solution for our code review tools, rewriting portions of our in-house CI/CD solution for performance improvements and compliance, and assisting with the rollout of observability tooling.

Technologies used: Python, Go, Ansible, Terraform, Honeycomb, Raygun

• Worked on migration from Pants targets to Bazel. Specifically worked on implementing Javascript targets, fixing Python test targets for production, and designing and implementing a BuildBuddy implementation that processes every single build at Oscar and surfaces helpful information to developers.

Technologies used: Bazel

Working on optimizations and next steps for Autopandas.

https://techcrunch.com/2017/11/01/accel-launches-uc-berkeley-mentorship-program/