Kirkland, Washington, United States
I have 9+ years of experience. Spent most of my career (6+ years) working in Microsoft Azure Networking / Load Balancer, a ring-0 highly available, geo-redundant and high performance distributed networking component. I’ve been a TLM for 2.5+ years, leading a team of 4 members. I own multiple control plane services and customer facing load balancing features (cross region load balancer, outbound SNAT, gateway load balancer, etc). Additionally, my team owns numerous quality and supportability initiatives. I’m also the service admin managing internal monitoring and logging resources, automations and data pipelines. I'm self driven and enthusiastic about driving innovative projects forward, while placing high importance on clarity, simplicity, quality and end-to-end ownership. I enjoy developing tools to streamline workflows, sharing knowledge and helping people to grow.
Infra - Shard Manager
Lead and mentor a team of 4 members, own feature definition, prioritization, design, delivery and maintenance. My team owns multiple control plane services, various customer facing load balancing scenarios and multiple quality and supportability initiatives. I end-to-end led and launched Global Load Balancer feature to GA, providing cross region redundancy, high availability, security and ultra low latency.
Own upper stream control plane of Gateway Load Balancer, an innovative patented project, facilitating a transparent insertion of network virtual appliances (NVA) into a cloud computing system. Conducted major refactoring that achieved seamless, backward-compatible goal state migration. Launched a novice probing datapath leveraging VxLan, removed dependency on low-performant Raw Socket, removed dependency on unreliable host node SNAT. Resulting in 50 % reduction in CPU load, fully eliminated transient failures during failover. Developed a light weighted in-house C# raw packet parser to enable IPv6 probing, contributing IPv6 load balancing GA. The parser is easily extensible, achieving an agile feature adoption for probing service.
In-depth knowledge of Azure networking & troubleshooting. Implemented canary incident auto enrich & auto triage on top of SREBot, built test framework to query historical data to test triage accuracy. Reached a 80% overall accuracy, drastically improved canary incidents TTE / TTM. Participated in multiple regions buildout, built automation tools to validate buildout correctness, saving days of human toil per buildout.
Care Everywhere team dedicated to clinical data exchange between organizations. Scenario owner of patient link / unlink, link forwarding, contact move, merge / unmerge.