* Researching and developing large-scale transformer foundation models for time-series forecasting and anomaly detection, scaling advanced capabilities across Google products.

* Developed and deployed a custom deep learning model for network traffic forecasting, significantly reducing congestion in Cloud Data Center Networks.

* Led ML Engineering for Ads Campaigns forecasting; developed novel deep learning models achieving the single largest quality improvement in the project's history.

* Launched multiple ML models across Google Ads and Google Ads Manager, powering downstream products that help advertisers and publishers optimize key performance metrics.

* Spearheaded the full-cycle development of high-performance distributed SVMs and kernel methods, integrating the library into TensorFlow.

* Improved Web table classification, directly increasing structured snippet coverage in Search and boosting user engagement at scale.

* Led design and development of a pipeline scaling category assignment fixes for a large corpus of Google Maps businesses.

* Unblocked team launch targets for precision and recall, resulting in a substantial increase in daily user impressions for affected listings.

I was a member of the Security and Cryptography group at UCSD. My main research focus was on understanding the use of hard learning problems in the construction of cryptographic primitives.