Senior Data Scientist

• Mar 2023 - Present

Applied Scientist

• Jun 2022 - Mar 2023

- Account Integrity team, working on proactively prevention of the phishing attacks targeting Amazon customers - Core Project: Devise end-to-end phishing websites detection ML systems to detect phishing attacks before customers' accounts are compromised - Machine Learning Research Areas: Multi-modality modeling, natural language processing, semi-supervised learning, clustering - Strengths: Formulation of business ideas into scientific problems, dedication and caringness for customers - Account Integrity team, working on proactively prevention of the phishing attacks targeting Amazon customers - Core Project: Devise end-to-end phishing websites detection ML systems to detect phishing attacks before customers' accounts are compromised - Machine Learning Research Areas: Multi-modality modeling, natural language processing, semi-supervised learning, clustering - Strengths: Formulation of business ideas into scientific problems, dedication and caringness for customers

Applied Scientist Intern

• Jun 2021 - Sep 2021

Developed semi-supervised learning models to correct noisy labels, and resulted in improved performance of Amazon account take-over detection model Developed a new data pipeline for offline automatic data labeling, as an alternative mediation to unblock online data warehouse failures Developed semi-supervised learning models to correct noisy labels, and resulted in improved performance of Amazon account take-over detection model Developed a new data pipeline for offline automatic data labeling, as an alternative mediation to unblock online data warehouse failures

Applied Scientist Intern

• Jun 2020 - Sep 2020

· Designed and developed order delivering defect forecasting system with tree-based algorithms (Random Forest, Isolation Forest, and XGBoost) for Amazon FLEX service · Applied SHAP method for explicit machine learning model explanation; successfully identified new useful features, and revealed these features' novel impacts on real-world scenarios · Designed and developed order delivering defect forecasting system with tree-based algorithms (Random Forest, Isolation Forest, and XGBoost) for Amazon FLEX service · Applied SHAP method for explicit machine learning model explanation; successfully identified new useful features, and revealed these features' novel impacts on real-world scenarios

Software Development Engineer intern

• Jul 2019 - Sep 2019

Developed an adversarial image detection application for Intel’s Edge Compute mini PCs, to protect customers’ deep learning model inference accuracy from contamination Developed an adversarial image detection application for Intel’s Edge Compute mini PCs, to protect customers’ deep learning model inference accuracy from contamination