Ads Marketplace (Ad Load/Insertion, Auction, Responsible Monetization)

• Designed and implemented five variations of the Auto Refinancing UI to support the Multi-Armed Bandit (MAB) algorithm, enabling dynamic testing and optimization of user experience.

• Integrated and deployed the MAB algorithm, enabling automated traffic distribution across the redesigned Auto Refinancing UIs and ensuring data-driven decision-making for future enhancements.

• Built a monitoring tool in Looker Studio to analyze user behavior and performance metrics from the MAB experiments, enabling continuous refinement of the Auto Refinancing UIs based on real-time insights.

Intensive Career Accelerator Program

• Selected as 1 of 292 out of 1300 applicants as a finalist for a fellowship that matches high-achieving Black and Hispanic students with top tech companies in the Bay Area/Silicon Valley.

• Proposed and developed a Machine Learning recommendation system utilizing NLP and Ontologies using Python.

• Improved contextual knowledge understanding and matchmaking by utilizing skill ontology to construct novel modified bag-of-words model, which lead to 65% increase in accuracy.

• Used Protege for Ontology management, NLTK for NLP, and Spark on AWS to run the models.