• Led a team of up to 8 engineers, onboarded 5 new engineers, and mentored 2 engineers. Planned new projects, set technical roadmap, and worked with product manager on the product roadmap.

• Set team objectives and created KPI dashboards using AWS QuickSight to track progress based on millions of data points, which convinced key stakeholders to prioritize reliability.

• Defined technical plan for reliability improvements and spearheaded development, which increased data pipeline reliability from 87% to >99% across the board, reducing time spent on support tickets and bugs.

• Managed cross-functional projects with the data science team resulting in streamlined model development that saved $500,000 dollars a year and reduced the time to deploy models.

• Developed media annotation pipelines that ingested up to 100,000 videos per day using Python, AWS Lambda, Step Functions, SQS, DynamoDB, MySQL, and Snowflake. Built accompanying service integration APIs in Typescript.

• Architected 2 new API integrations with commodity ML providers and integrated 7 new proprietary models.

• Introduced AWS Sagemaker into the media annotation pipeline reducing time-to-value for new proprietary models from weeks to days.

• Diagnosed and fixed scaling issues with distributed systems, which helped hit the goal of doubling capacity every year.

• Launched a serverless analytics pipeline for a new flagship product that handled 10 million rows of data per day.

• Championed a testing culture by introducing end-to-end testing for data pipelines and directing work to increase unit test coverage from 23% to 63%.

• Sped up a key ad analytics ingestion pipeline by 35% by maximizing I/O usage.

• Manager of a student-run café.

• Working with my comanagers, doubled revenue by expanding offerings and lowering costs while fully paying off debts totaling over $20,000.

• Developed new systems to speed up our hiring process, centralize our finances, and reduce time spent doing inventory.

• Redesigned a provisioning API for an internal inventory management system from Perl to a modern stack (Node.js, Express.js, MySQL) while retaining backwards compatibility

• Suggested and implemented improvements to the code base for the larger inventory

management system including refactoring to reduce complexity, adding monitoring metrics and end-to-end tests, and separating application layers