Led data and model understanding in order to infer metadata required for compliance projects, covering over 20,000 datasets and 3,000 models. This supports compliance projects such as DMA and the EU AI Act. Set and implemented a roadmap for a team of three.

Worked cross functionally with the PM organization and legal team to develop and implement technical strategy for EU AI Act and GPAI model identification, implementing systems to automatically classify 80+% of models, and saving engineers hundreds of hours.

Led the end-to-end cross-functional development of sensitive data detection pipeline for NAS files, cataloging and sampling hundreds of millions of files using Spark, SQL, and go cron jobs. Won quarterly spot bonus for collaboration.

Fulfilled support requests and built features for TensorFlow model training used by our company-wide machine learning platform.

Migrated the cross-team TensorFlow layers library to keras. Built pytest regression tests and manually validated production model training workflows after migration.

Led post mortem after a model performance regression and built regression testing for model quality across the ad-ranking team. Became owner of all integration testing for model training.

Designed and implemented the recommendation system from scratch for all Creative Tools, working with partner teams and running a roadmap for a team of two along with launching the initial product.

Built data pipelines to generate lists of content candidates for retrieval using AirFlow.