Experience

Honda

• Japan
• Motor Vehicle Manufacturing
• 700 & Above Employee

• Capstone Project Researcher

  • Sep 2023 - Present

  • Developed non-linear models with DCM (Complexity Scores) for precise cost predictions of Honda automobiles • Crafted a Python package that utilizes K-means clustering to categorize DCM values into distinct groups • Developed non-linear models with DCM (Complexity Scores) for precise cost predictions of Honda automobiles • Crafted a Python package that utilizes K-means clustering to categorize DCM values into distinct groups



Amazon Web Services (AWS)

• United States
• IT Services and IT Consulting
• 700 & Above Employee

• Software Engineer Intern

  • Jun 2023 - Present

  • Designed a Java based demand forecasting package using the Triple Exponential Smoothing time series model; deployed via AWS's CDK (Cloud Development Kit) on EMR Serverless with Apache Spark • Stored model parameters for items in DynamoDB and archived forecast results in the targeted S3 bucket • Established an automated AWS pipeline with Lambda and Step Functions; collaborated with the data science team to devise a tuning strategy for multi-granularity forecasting • Enhanced SparkSQL and JavaRDD to process 200 million records during load tests; achieved a 33% speed improvement over the standard Amazon Forecast solution for Amazon Pharmacy datasets Show less



TransUnion

• United States
• IT Services and IT Consulting
• 700 & Above Employee

• Industry Practicum Researcher

  • Sep 2022 - May 2023

  • Identified and assessed data drift using critical metrics such as KS (Kolmogorov-Smirnov) and PSI (Population Stability Index) for personal loan and home equity datasets • Refined the base XGBoost model by integrating eDGBT (evolving Distributed Gradient Boosting Tree) algorithms with a regrowth technique • Developed a user-interactive auto-training pipeline for efficient model retraining, ensuring robust adaptability and accuracy • Enhanced model performance in the final deployment by achieving 23% increase in the KS score to minimize drift effect Show less



China Unicom

• Telecommunications
• 700 & Above Employee

• Data Scientist Intern

  • Jun 2021 - Aug 2021

  • Implemented pipeline with Python to analyze AI text proofreading demand; Utilized RandomForest models from the sklearn and harnessed multiple visualization libraries, including matplotlib, seaborn, and Plotly • Collaborated with machine learning Engineers to optimize neural network models in the context of official title hierarchy, leading to an 18% increase in baseline accuracy • Implemented pipeline with Python to analyze AI text proofreading demand; Utilized RandomForest models from the sklearn and harnessed multiple visualization libraries, including matplotlib, seaborn, and Plotly • Collaborated with machine learning Engineers to optimize neural network models in the context of official title hierarchy, leading to an 18% increase in baseline accuracy



杭州电魂网络科技有限公司 Electronic Soul

• China
• Computer Games
• 1 - 100 Employee

• Software Engineer Intern

  • Jun 2020 - Aug 2020

  • Developed user-friendly front end using JavaScript/HTML through Vue.js to help view and update player profile • Generated MySQL relational schema to allow users keep track of game history and account condition • Implemented REST APIs with Flask in Python to retrieve and cache user information • Wrote pytest with 90% code coverage to ensure program worked correctly in processing data • Developed user-friendly front end using JavaScript/HTML through Vue.js to help view and update player profile • Generated MySQL relational schema to allow users keep track of game history and account condition • Implemented REST APIs with Flask in Python to retrieve and cache user information • Wrote pytest with 90% code coverage to ensure program worked correctly in processing data



University of Illinois Urbana-Champaign

• United States
• Higher Education
• 700 & Above Employee

• Undergraduate Course Assistant

  • Jan 2020 - May 2020

  • Served as the course assistant for Intro to Programming CS course with a size of 500 students. • Hosted weekly office hours to help students solve homework and machine projects problems via Java. • Guided students through basic computer concepts and assisted students to debug in the Android Studio environment. • Served as the course assistant for Intro to Programming CS course with a size of 500 students. • Hosted weekly office hours to help students solve homework and machine projects problems via Java. • Guided students through basic computer concepts and assisted students to debug in the Android Studio environment.