firstname.lastname@example.org • https://www.linkedin.com/in/sunhyoung-han/• CA, 92131 • 858-***-**** Key Competencies
Dedicated and accomplished leader with extensive experience in machine learning, research and development, fraud detection/prevention, and risk management. Stellar record in devising IT solutions, marketing analytics, price optimization, process improvement, and business development. Proven track record of fraud control, business modeling, customer lifetime value management, and budgeting and forecasting. Seasoned in credit line allocation, credit risk management, vendor management, algorithmic enhancement.
• Credit Risk Detection
• Artificial Intelligence
• Machine Learning
• Strategy Development
• Database Management
• Project Management
• Fraud Detection/Prevention
• Business Intelligence
Zebit, San Diego, CA
VP Chief Analytics Officer 11/2018 to Present
Collaborate with senior leadership team and transform the company from expert knowledge base to data driven decision making. Lead advanced analytic capabilities and solutions supporting all business functions including risk management, marketing analytics, data analytics operations, software vendors, and price optimization. Administer third party data strategy, performance evaluation, adoption, analytic vision, strategy, and roadmap to achieve business success. Utilize Machine Learning to identify pain points in business and process improvement opportunities. Establish and execute adaptive learning loop to measure, monitor, and enhance business decisions. Assist and build Data Science teams to deliver results at the speed of the business. Apply vendor partnership strategy to ensure timely delivery of analytics solutions. Supervise data analytics, custodianship, and infrastructure to ensure alignment with the Data and Analytics department, avoiding conflicting activities, and availing the most efficient data analytics insights across the business. Oversee cross-functional data governance while simultaneously ensure adoption and adherence to data quality and process governance in the relevant collaborating departments.
● Doubled the revenue by keeping bad debt rate marginally increased by applying machine learning models in production and handled targeted fraud attack with Fraud model in production.
● Amplified profitability by 50% by building comprehensive business model ranging from customer acquisition cost optimization, risk management, and price optimization.
● Increased consumer's LTV (lifetime value) by 30% with internal policy optimization.
● Reduced fraud operation team's manual review by 50% with automatic fraud risk model.
● Improved credit loss forecasting accuracy by 20% by introducing Bayesian framework.
Head of Risk Analytics 02/2018 to 10/2018
Directed development and implementation of risk models to reduce risk of payment defaults starting from Zebit’s the first ML based model. Enhanced data cost for customer acquisition, collection strategy, third party data sources, customer’s credit line allocation, and increased strategy over consumer’s life cycle.
● Established marketing budget allocation optimization and enhanced production.
● Reduced payment default rate by 10% with machine learning deployment in production and customer acquisition cost by 30% from successful partnership with selected third party vendors.
● Obtained 10% additional customer acquisition by optimal marketing budget allocation and decreased customer payment risk by 10% with optimal credit line allocation. ID Analytics Symantec, San Diego
Sr. Principal Scientist Principal Scientist Sr. Scientist Scientist 02/2011 to 01/2018 Headed ID Analytics’ major CreditOptics models development and release which predict consumer’s credit risks when consumers apply for credit card, bank account, auto finances, mortgages, and wireless services. IDA solutions are used by, 6 of the top 10 financial institutions, 4 of the top 5 wireless carriers, and 4 of the top 5 retail card issuers. Contributed for algorithmic enhancement, assisted in developing model governance strategy setting for the company, and led multiple projects involving 3-9 Ph.D. scientists and software engineers. Applied convolutional neural network in fraud/credit risk prediction modeling and used CNN as data analysis tool to enhance tagging and data visualization. Successfully applied CNN to new product development and contributed the architecture of CNN which is customized for credit assessment modeling. Established an infrastructure for parallel batch queueing system for Analytics department.
● Amplified risk model accuracy by 10% by introducing Convolutional Neural Network in the credit risk industry and built hybrid model architecture exploiting different strengths of machine learning algorithms.
● Built disparate impact aware modeling technology which reduces disparate impact by 30-50% without sacrificing model performance.
● Built consortium and custom credit/fraud models for bank, retail, and wireless industries for IDA's B2B customers, which provided 10-50% better performance than its predecessors.
● Achieved first place in the bake-off competition in alternative credit score space from a model powered by uniquely designed "geo trend" variables.
● Improved credit/fraud model performance by 10% with risk-based feature transform. The suggested transform also contributed to reduce training time by 40%.
● Built probabilistic model to check validity of SSN with SSN randomization policy from SSA. The model achieves 20% better accuracy.
Graduate Researcher Statistical Visual Computing Lab, Intern Qualcomm, Sr. Researcher LetsVision Co., Research Engineer Hynix Seminconductor Inc.
Doctor of Philosophy, Electrical and Computer Engineering University of California, San Diego, CA
Master of Science, Electrical Engineering
Yonsei University, Seoul, Korea
Bachelor of Science, Electrical Engineering
Yonsei University, Seoul, Korea
Sunhyoung Han, Nuno Vasconcelos, “Object recognition with hierarchical discriminant saliency networks” Frontiers in Computational NeuroScience, 8:109. doi: 10.3389, 2014 Sunhyoung Han, Nuno Vasconcelos, “Biologically Plausible Saliency Mechanisms Improve Feedforward Object Recognition” Vision Research, vol, 50(22), pp. 2295-2307, October 2010. Dashan GAO, Sunhyoung Han, Nuno Vasconcelos, “Discriminant saliency, the detection of suspicious coincidences, and applications to visual
recognition”, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol 31(6), pp. 9891005, June 2009