Brian Hu Email & Phone Number

Seattle, Wa, Us

 Developed “Blueprint”, an innovative data-driven organizational design and inspection prototype, currently leveraged by 300+ high profile active customers across Amazon (VPs, Directors and HRBPs) for organizational design and planning purposes; Developed organizational design related science models that are currently implemented in the “Blueprint” prototype:• A graph neural network deep learning model to measure attrition rate given organizational structures• An innovative graph-based method to measure organizational complexity and link the complexity to promotion and organizational longevity Developed the development of recommendation system (using deep learning, matrix factorization and other modeling frameworks) to recommend the best HR products to 1million+ hourly Amazon Associates with the ultimate purpose to reduce attrition; Developed a Natural Language Processing model-based prototype to determine the matching job family and level for potential employee candidates; the prototype was used in ~20 acquisition and merger deals and saved around ~10,000 hours analyst time annually; Developed a Natural Language Processing/Large Language Model based chat-bot that automate answering questions related to internal HR products; the prototype is expected save product managers ~50,000 hours annually. Lead the development of a Natural Language Processing/Large Language Model (“LLM”) based chat-bot that automate answering questions related to Paycom products; the tool is currently leverage by 3000+ internal employees and estimated to save $10MM annually; Lead the continuous monitoring and research effort to improve the model suite, including:• Set up model evaluation metrics to guide the various choices of model options• Search for additional data sources • Implement methods in cutting edge research papers to potentially enhance the model• Identify and research on strategies on migrating model risks

San Antonio, Texas, Us

DATA SCIENTIST MANAGER (CAO DEPARTMENT)  Led a team of 3 scientists for the development and life cycle management of supervised and unsupervised machine learning models:•Unsupervised Natural Language Processing model suite to aid the contract review for procumbent/legal department (Estimated saving: $10+ MM/year)•Supervised Sensitive Data Management model to categorize the sensitivity level of internal files (Estimated saving: $2+ MM/year)•Supervised AI chatbot model as part of IT ticket management tool (Estimated saving: $3-4 MM/year) Develop and successfully launch model development control checklist by working closely with executives and project managers and other shareholders; Participate in a panel overseeing USAA data scientist job governance; Coach two junior data scientists to successfully pass review exams for promotion; Peer review multiple machine learning/time series forecasting models used across enterprise;  Lead Pulse channel to facilitate the communication between the management team and the employees, while keep monitoring repeated feedback; Successfully completed USAA Manager Candidate School training.LEAD MODELING ANALYST (CFO DEPARTMENT) Facilitated in the development of first-line model development policy/standard (The policy/standard were successfully approved by Board of Directors for governing all enterprise models at USAA); Led the development of multiple time series forecasting and machine learning models (models were well acknowledged during regulator’s on-site examination):  Developed a new model documentation guideline to help successfully close audit findings; Facilitated the executive review and committee approval of the guideline; Collaborated with CFO data and IT team to optimize the data pull process;  Worked closely with executives and project managers in identifying and eliminating interim gaps in model building processes.

SR. MODELING ANALYST (CFO DEPARTMENT)  Lead a modeling team of junior statisticians and business analysts in building revenue/fee income forecasting and operational risk simulation models for bank stress testing purpose; Collaborated with model validation team to refine and improve model risk management processes, such as remediating model deficiencies, establishing model performance monitoring, sensitivity and buffer analysis framework; Developed detailed documentation of process and procedures related to model development and implementation process; Worked with Scenario Development Group in developing stress test scenarios used company-wise; Participated in multiple rounds of model results business unit challenge by the bank executives; Collaborated with Financial Planning and Analysis (FP&A) and Treasury departments to improve Net Interest Income forecasting process.

Master'S Degree, Computational Finance And Risk Management

Doctor Of Philosophy (Ph.D.), Biotechnology