Project Usher - As lead backend developer, designed, architected, monitored, and deployed extensible HATEOAS inspired-REST backend for COVID monitoring using biometric facial recognition

• Close integration work with the frontend team across multiple customers

• Interconnected Java microservices using PostgreSQL, Kafka, Redis, Amazon Simple Workflow Service (SWF), and AWS Lambda in FedRAMP environment

• Mentored and onboarded junior engineers to AWS SWF and Lambda Functions

• Python end to end tests

• Designed terraform to interconnect and deploy AWS Lambda Functions with SWF microservices

• Modified Jenkins pipelines as needed to enable deployments

Logistics Industry, AWS cloud, Java, Spring, Microservices, PostgreSQL

• Installed Sleuth tracing framework to trace HTTP requests throughout dozens of microservices

• Implemented Graphana alerts to monitor services

• Implemented PostgreSQL-backed microservice on AWS to store and encrypt customer credentials

• Created Java microservice to poll for shipment updates and write them to a Kafka topic

Packed Storage- Updated disk storage solution improving performance via multiple objects per file

• Implemented unclean shutdown and corruption detection regression tests in Python

• Corrected clean and unclean shutdown handling and memory management issues

• Led effort to handle data integrity issues on a more granular layer improving storage reliability

• Developed throttling of background disk compaction activity to mitigate performance variability

• Created metrics to allow for proactive monitoring of compaction activity by support

Fanout- Optimized storage of multiple copies of objects and minimize network usage

Designed, led, and completed majority of implementation

• Handled impacts at all layers of the stack

• Used gathering writes and batching to improve performance

• Gained experience with changing requirements

• Took initiative to work with the support team to diagnose problems and prioritize and design solutions

Disk Lifecycle Management v2–Updated notification interface for new set of DLM events

• Consumed new HTTP and RabbitMQ API to communicate disk state changes between processes

• Adapted Python regression tests for new API