Hire Chao L. - Data Scientist

* Collaborated closely on engineering team and UX team to initiate and formulate hypotheses, conduct A/B tests, evaluate experimental results to help product team on decision making. * Designed metrics for evaluating online recommendation & search algorithms and implemented data pipelines for A/B testing. * Conducted comprehensive audits of data quality for the launch of new website platforms and various business initiatives.

* Managed risk rule stacks and utilized ML techniques to detect real-time fraud attacks for subscription-based products. * Built tree-based models for fraud detection, implemented rule stacks from scratch, and successfully onboarded several new business to risk platform. * Conducted A/B tests to evaluate the performance of different risk models, implemented strategies for smooth transitions among online models, and closely monitoring their performance.

Developed a web application www.givamekiva.com to predict the success rates of loans on Kiva, a micro-funding platform, so that users can know if their loans are at risk of not being funded. Performed various topic modeling methods including LDA, TF-IDF to extract underlying topics from personal stories and tags on kiva campaign websites. Trained a Light GBM classifier and obtained a roc-auc score of 0.86. Interpreted the predictions of the classifier using LIME and gave feedbacks to users on how to improve the content of the campaign messaging in order to increase the chance of loans being successfully funded. Web application was built with Flask and deployed on AWS.