United States
I serve as a startup founder and CEO. Previously, I was the project lead and owner for the next generation route planning engine in DiDi. The team consists of top notch machine learning engineers, software developers and QAs. I also led the route planning (RP) team and delivered the RP service in South American countries. Besides, I mentored the juniors and helped them reach beyond what they think could be possible. Before I joined DiDi, I was the data science lead in Trend Micro and delivered the machine learning solutions to detect malicious scripts and protect the Macintosh. I also led the team for the Trend Micro Global AI Contest and won the 1st place among 91 teams in the CDC office. When I was in graduate school at Drexel University, I was the leading contributor to the ReTouch Lab on haptics which is a branch of virtual reality. One of the paper we published was in the best paper finalist of the IEEE Haptics Symposium 2016. I was also the teaching assistant in the ECE department and mentored hundreds of students around the world.
Route Planning for Domestic Business • Design, develop and deliver as project owner the next generation route planning engine powered by the cutting-edge machine learning solutions. • This project has been delivered and launched in DiDi's app to assist tens of thousands of drivers in domestic China. Route Planning for International Business • Managed the route planning (RP) team including machine learning engineers, software developers and QAs to deliver machine learning based RP service in South America. • Increased the exact similarity ratio by 1.09% of the top 1 route (Top 1 Sim100) compared to Google which is 0.89% • Decreased the Broken Experience Rate (BER) by 1.08% versus 0.85% by Google
Team lead for Trend Micro AI Contest (CDC office rank: 1/91): Designed agent to play Hearts based on Monte Carlo Tree Search (MCTS) which can shoot the moon and beat rule based robot crafted by human. Project lead: Delivered graph neural network (GNN) models for threat prevention and device attribute tagging Delivered machine learning models to detect bruteforce attacks and lateral movement attacks under AWS environment
Project lead: Detect malicious scripts including PyCryptoMiner that steals bitcoin Results: Delivered machine learning solutions for Trend Micro Antivirus Mac to participate AV Test and AV Comparatives and both of which ranked No. 1 among 11 global cyber security vendors including Norton Security, Kaspersky and Bitdefender
Capstone project: Evaluate Amazon Book Reviews ( https://github.com/Phil-U-U/Evaluate-Amazon-Book-Reviews ) Case Studies: Sale Price Prediction of heavy equipment at auction; Churn Prediction and User Retention Analysis; Recommender System modeling and User Behavior analysis; Fraud Detection Topics covered: Regression (Linear Regression, Logistic Regression, Regularized Regression); Supervised Learning (K Nearest Neighbors, Decision Trees, Bagging and Random Forests, Boosting, Support Vector Machine); Unsupervised Learning (K Means and Hierarchical Clustering, Dimensionality Reduction and Non-negative Matrix Factorization); Big Data Engineering (Graph Theory, Map Reduce, Spark, AWS and Parallelization, Data Visualization)
Research Topics: • Estimate Social Network Characteristics via Random Walk Sampling • Grasping Data Analysis and Haptic Working Memory Modeling • Load Consumption Flattening for Power System
Teaching assistant for the following courses: • Programming for Engineers (Python) • Fundamentals of Systems I & II • Digital Logic Design • ECE Laboratory