I am a Senior Software Engineer at Klaviyo based in New York City. Most of my work is ensuring the scalability and stability of our Campaign sending pipeline, which regularly sends over a billion emails a day, double on Black Friday.
Factored out campaign recipient generation in sending pipeline to allow for independent scaling, multiprocessing, task impotency, and giving users the ability to generate up to 15,000,000 recipients at send-time for campaigns scheduled in the future.
•
Created internal querying service abstraction to reduce downstream modifications and exceptions, allowing us to move out of the mono-repo while preserving availability.
•
Used Amazon Database Migration Service to consolidate campaign channels (email, SMS, and push) to allow for multiple-channel Campaigns and reduce the number of database calls by a 67% from end-users looking to see overviews.
•
Migrated a-sync task processing from RabbitMQ to a Pulsar-based internal tool, dogfooding it
•
Used AWS CLI and Blue/Green deployments to partially automate engineering-wide MySQL 5.7 -> MySQL 8.0
•
Assisted onboarding of new team members by helping develop new documentation to reduce tribal knowledge.
•
Helped recruiting efforts at Cornell and led technical interviews for 100+ SWE candidates-
Used Redis, React, TypeScript, and Django's ORM to create a banner displaying a cached value of how users compare in open rates and click rates to other users in their market category.
•
Identified and mitigated bottlenecks and hardware limitations in the sending pipeline to allow us to prepare for Black Friday traffic, which typically was double the previous year's traffic.
•
Split out fields on campaign model to reduce model bloat and helped direct transition of over 1,000 unit and integration tests that need to be transitioned from old deprecated model
Worked with team product manager and large customers to define and create new public API endpoints to allow the retrieving email open and (link) click rates in a secure and RESTful way
•
Redeveloped the email quarantine algorithm to reduce the number of private email addresses incorrectly not receiving emails by approximately 80%.
