Developed AI/ML-driven features in Cloud Cost Management and Optimization SaaS productsAgent building, Retrieval Improvement and Evaluation, Anomaly Detection Service, Text Classification

- Posted news, technology, personal & family practical tips, and local events & resources related to sustainability, climate changes, and eco-friendly innovations across three social media platforms

The goal is to ensure that AWS has enough capacity to meet customer demand while improving energy usage effectiveness in global data centers. - Built and productionized machine learning models to build scalable tools/systems, driving how racks are ordered, allocated, and deployed in global data centers. - Coached team members in different job families (data scientist/data engineer/software engineer/business intelligence engineer) to improve the processes around building data pipelines, deploying model updates and monitoring model performances. - Experienced interviewer, with 70+ interviews performed at Amazon.

B2B Payment Products, Apr 2019 - May 2020EU5 Payment Products, July 2017 - Mar 2019Amazon Currency Converter, Oct 2016 - June 2017- Developed propensity models, time series analysis, A/B testing, and BI solutions for global payment products.- Managed data science sprints and prioritization processes with 20+ product/marketing/finance stakeholders. Designed basic SQL and statistics training sessions to empower stakeholders for self-service analytics.- Set up a monthly data science learning series within the Consumer Payments org; onboarded new data scientists and senior business analysts.

- Built models to predict loss risk and estimate policy lifetime value for 15 states.- Designed optimal SAS raters for the new auto insurance policy in California, which reduced the working time of business analysts by 20%. - Coaches new hires, educated peers with advanced statistics knowledge and standardized tools.

- Partnered with US public school districts to rate 10k+ teachers' performances using student assessment records.- Built fixed-effect and random-effect models to measure teachers contributions to student learning growth.- Developed and standardized SAS Macro codes for data cleaning, validation, reporting, and quality checks; applied parallel ideas into SAS programming, which reduced the computing time by 40%+.