Coco(Keying) Chen Email & Phone Number

Toronto, On, Ca

Toronto, Ontario, Canada

- Implemented end-to-end pipelines in DBT to map shopper interactions across various business partners, enabling analytics stakeholders to conduct bidding recommendation experiments, led to a 15% increase in vendor bids and a 10% increase in revenue.- Included data quality monitors in pipelines to alert teams about issues eg. data freshness, schema changes, and abnormal data volume, resulting in proactive measures to mitigate impact on 100+ business stakeholders.- Developed lightweight UI for dbt seed and backfill operations utilizing Airflow’s intake feature. This automation significantly enhanced the team’s KTLO efficiency and reduced ongoing maintenance efforts.- Championed development of On Call Runbooks to document on call best practices eg. redeployment of Trino clusters and standardize root case analysis approaches, reducing similar incident time to resolution by 70%.- Overhauled the Python API ingestion pipeline using vendor’s REST API endpoints, processing over 100k call recording data per day to mitigate company’s loss of functionality of compliance reviews while maintaining vendor relations.- Refactored the ingestion pipeline bottlenecked by memory and rate limit constraints using Python and Airflow which results in reduction of task runtime by 50%.- Improved data pipeline infrastructure by consolidating common logics into generalized DBT modules to support seamless onboarding of additional lines of business. This led to reduced internal development time, improved code standardization, and supported company-wide compliance initiatives.- Hosted presentations and knowledge-sharing sessions to upskill a team of 38 data engineers, datascientists, and business analysts on Trino best practices to improve query performance and reduce cluster usage costs.

Toronto, Ontario, Canada

- Developed predictive models for hourly wind and solar resource outputs using a stacked model of Neural Networks, Linear Regression, Random Forest, and AdaBoost in Scikit-learn and PyTorch. Contributed to the development of two electricity demand forecasting products, enhancing the accuracy of renewable energy supply management.- Supported the development of electricity demand forecasts for the Annual Planning Outlook, improving predictive accuracy and contributing to long-term (20-year) insights about Ontario’s electricity system.- Automated bi-weekly team progress reports and the import of industrial raw data, reducing repetition and improving team productivity by 30%.