Celine Liu

Led a team of data engineers to successfully complete complex projects on time and within budget, projects include but not limited to:

• Automated manual and repetitive data tasks using Python or HSQL, saving time during the analytical process significantly
• Content Assurance: Set up data pipeline for CA metrics and dashboard on internal data platforms for U.S video (views) boosting programs, made sure the program is compliant with US Data Privacy and Security Operation policies; validated metric values and numbers created using two separate data generation mechanisms, investigated and closed gaps in a timely manner
• Data Loss Prevention: Fetched weekly reporting data from the 3rd party UI (CSV file) to be used as the data source, built data pipeline for DLP metrics calculation and dashboard
• Software Assurance: Extracted relevant raw data from 3rd party, transformed & loaded data to internal data platforms and built data pipelines to generate daily monitoring reports and calculate SBOMs (Software Bill of Materials) metrics such as # of submitted/approved/failed, visual queries and summary dashboard

Built strong relationships with external partners such as vendors or consultants to facilitate collaboration on joint projects and initiatives

Discovered stories told by data to present information and provide insights to leadership team

• Developed innovative teaching methods tailored to diverse learning styles for course 'Introduction to Artificial Intelligence', enhanced student comprehension and success rates
• Taught 'Smart E-Commerce: Recommendation Systems: Design and Application' (lecturing, real examples and online coding), used different learning modes and types of technology to engage students in achieving their learning outcomes
• Developed diversified course curriculum, tests and educated 'Smart E-Commerce: Recommendation Systems: Implementation and Maintenance' (lecturing, real examples and online coding).

• Built a job recommendation system for the users on our job search platform (utilized RoBERTa for embeddings), increased job-resume matching efficiency by over 30%
• Developed a tool to provide personalized career advice and suggest skill development paths based on a user's profile and career goals
• Mentored junior team members, contributing to their professional growth and development as successful data scientists.

• Predicted probabilities for their next status for client's loan products such as CRE, C&I, Mortgage loans, etc, using a survival analysis combined with machine learning techniques (random survival forests)
• Experimented with other machine learning algorithms, for instance, LightGBM and Multinomial logistic regression and compared their out-of-sample testing results
• Collaborated with cross-functional teams to ensure seamless integration of models into existing systems and workflows.

• Built loan level quantitative model to predict default possibilities for client's U.S Residential mortgage portfolios and deployed CCAR models for production
• Conducted in-depth analysis of model performance and identified inherent relationships between the behaviors of client's U.S Residential mortgage loans and their risk drivers.
• Liaised with client and management team to better understand their needs and recommend appropriate technical solutions.

• Develop PPNR models for fixed income, equity derivatives and global short-term market
• Designed and built logic-based computation tool to forecast revenue and expense/cost under different stress scenarios.

• Built quantile regression and OLS regression to predict PPNR for commercial & retail loans, deposits, and cash management fees, performed back-testing
• Performed sensitivity analysis under different stress scenarios for CCAR stress testing
• Validated modeling methodology for residential mortgage prepayment prediction.