1. Build distributed OpenTelemetry tracing to track the insights of an important API across different microservices.

2. Several optimizations on our Ray mapping pipeline to reduce processing time and unnecessary costs.

1. Make existing AR game server running on regional servers. Greatly reduced TTL for developers around the world.

2. Build a Python inference service to check the quality of scans uploaded. Design the service as a fault domain to prevent this service from impacting other important service.

3. Implementing content serving APIs for our localization service, use multi-thread programming for url-generation tasks to improve parallelism.

1. Designed and implemented a new revenue bookkeeping workflow in a Java-based stream-processing bookkeeping platform using Kafka and Flink. The new bookkeeping workflow is idempotent and resilient to reprocessing, which greatly reduced the accounting team’s repetitive work.

2. Managed all cross-team communications related to DashPass bookkeeping, including new feature initiatives, audit requirements, and quality support. Communicated effectively to reduce ambiguity.

1. Designed and implemented a React UI Rule Editor Framework, allowing developers to easily reuse components to build a rule editor website. Set up a test framework to enable automatic unit tests with a 75% webpage component coverage. Mentored an intern on a full-stack project based on this framework.

2. Designed and implemented a new labor planning workflow in our Java-based labor planning service. This involved converting labor assignment requirements into a generic constraint-solving problem. The new workflow helped reduce the time warehouse managers spent on labor planning.

3. Constantly improved the shipment tracking service, including data polling optimizations, progressive operational-excellence improvements, and flexible cloud resource stacks using AWS CDK.