* Researched, fine-tuned, and deployed neural translation models in 16 languages to generate billions of translations for our toxicity detections. * Led a talented intern team to research and create an LLM labeling system that measured performance of 16 different production models across 17 languages. * Created an LLM filtering system that improved precision of our production models in foreign languages by 20%. * Trained and deployed an identity hate and bias classifier for video game moderation that detected millions of player identity attacks. * Built grooming indicator models whose outputs were aggregated to detect real grooming alerts for our customers.

* Researched a novel neural network architecture to address challenging business problems and implemented it in Tensorflow and Python.* Engineered an end to end hierarchical event clustering pipeline that recognizes business process activities for a diverse range of client businesses. * Designed and implemented an unsupervised hierarchical labeling system that summarizes complex events and activities with intuitive labels. * Developed a self-supervised rule learning system that continuously provides event classification rules to clients.

* Created validation pipelines in Ruby and Python to build, calculate, and statistically analyze our Store/Usage Intelligence data.* Investigated data sources to identify usage intelligence trends and outliers.

* Predicted 1000’s of actionable events from terabytes of data by using a self-designed unsupervised learning approach to solve a challenging problem.* Designed and coded customer intelligence pipelines that find behavioral patterns from billions of mobile GPS data using big data and machine learning.* Created a validation system of statistical heuristics and Python code to animate different behavioral models and compare business trends in Google Maps.* Delivered Python programs that produce GPS traffic layers for the website.

* Successfully created a self-driving, collision evading quadcopter that dodges projectiles with partner Jeremy Crowley.* Awarded competitive Regent’s Fellowship as an undergraduate for research potential.* Led a team of high school interns to compete at the national Siemens Competition for robotics.

*Co-authored a research publication entitled "Ultra Fast Silicon Detectors."*Used C++ based simulation program Weightfield to quantitatively model silicon detectors. *Authored a 50 page research report in which theoretical framework of semi-conductor physics was outlined; original data obtained through Weightfield was presented; and algorithmic flaws in Weightfield's structure were identified--including comparative results demonstrating improved performance.