Experience

Yahoo

• Australia
• Online Media
• 100 - 200 Employee

• Senior Research Engineer

  • Mar 2022 - Present



Baidu, Inc.

• China
• Software Development
• 700 & Above Employee

• Senior Machine Learning Engineer

  • Dec 2020 - Feb 2022

  • Advertising optimization - Develop audience and content labeling for precise targeting in native ad system - Analyze issues from campaign targeting with proper solutions • Recommender System - Improve content recommender system with graph-based and DNN-based models on the retrieval and ranking stages • Advertising optimization - Develop audience and content labeling for precise targeting in native ad system - Analyze issues from campaign targeting with proper solutions • Recommender System - Improve content recommender system with graph-based and DNN-based models on the retrieval and ranking stages



Trend Micro

• Japan
• Software Development
• 700 & Above Employee

• Senior Data Scientist

  • Apr 2020 - Nov 2020

  • Research on end-to-end solution for static-layer threat detection - Image-based deep learning model with model interpretability in order to improve machine learning models on production - Analyze FN/FP samples with interpretable tools such as SHAP • Research on end-to-end solution for static-layer threat detection - Image-based deep learning model with model interpretability in order to improve machine learning models on production - Analyze FN/FP samples with interpretable tools such as SHAP



Vpon Big Data Group

• Taiwan
• Software Development
• 100 - 200 Employee

• Senior Data Scientist

  • Sep 2016 - Mar 2020

  • Build machine learning models and applications to provide audience labeling for precise ad delivery with third-party verification • Analyze user profile data and ad response data to predict user behavior and conversion probability • Develop Data Management Platform (DMP) - Data ETL pipeline from multiple source data using PySpark and Hive - Audience segmentation - Develop a microservice of exporting target audience to ad platforms via RESTful APIs using Flask and Docker - Implement crawlers to enhance audience insight report with simulated actions and APIs • Co-work with team members to build an interactive visualization system, TraVis, for efficiently mining travelers' activities from large-scale ad request data • Research on urban computing and trajectory analysis to benefit recommendation and prediction - Publication: "TraVis: An Interactive Visualization System for Mining Inbound Traveler Activities by Leveraging Mobile Ad Request Data. In Proceedings of the 28th ACM International Conference on Information and Knowledge Management", CIKM 2019. - Develop trajectory-based user lookalike model to optimize advertising strategies Show less