Alan Lee, Ph.D Email and Phone Number

Seeding and advising various early stage startups. Principal and Chief data, data science and AIML development. Heavy in data science and Gen AI development.

Rapidly hired a small team and rearchitected Nike's online search system with AI capability, delivering 140M in incremental revenue as a result of better search results.

Part of an exciting and high performing Data Science group that has quickly made breakthroughs at Kohl's. In a few years we have rapidly grown the core data science team, hiring experts in a wide range of areas including deep learning, countless regression techniques, tree ensembles, Bayesian statistics, and more. This team of currently of data scientists and data science engineers has solved many important problems at Kohl's and is being asked to take over more responsibilities across the entire enterprise. This team modernized AI/ML for many areas including recs, search, supply chain, marketing and other areas.Data Science and Engineering responsibilities in the group includes:-Online Recommendations and Personalization for Kohl's E-Comm assets - Research, Implementation, and Support-Search and browse algorithms-Advanced Marketing Data Science-Enterprise Wide Assortment Optimization -supply chain AI-Sales Forecasting-Visual Search using Deep learning(CNN) and traditional BOVW -Other adhoc projects analyzing sales performance, customer behavior, fashion trends, merchandising, etc to support critical decisions across the enterprise -Various Data Science Engineering challenges necessary to scale up big data machine learning solutions for deployment

Leading the Data Science division on exciting initiatives involving geospatial statistics, social network analysis, scaling Big Data algorithms on the cloud, traditional predictive modeling, advanced machine learning, and other data science related work. • Created, recruited, hired, and trained world class data science team from scratch• Rapidly architected and lead team to build comprehensive Big-Data Insurance fraud detection solution based on AWS cloud infrastructure using open source solutions(Hadoop,Redshift,Scala, etc)• Managed Full stack development team of developers and data scientists• Supported New Product Analytics Initiatives with fully functional prototype solutions including UI as well as critical in-house analytics projects such as companywide HR intelligence.• Personally responsible for architecting and prototyping most of the critical solutions • Received extremely positive feedback from customers and successfully created a working solution in months that had been eluding the company for years.• Built solutions based only on high level strategic objectives and extensive industry research including literature review searches in problem space• Solutions built include social network analysis, text mining, geospatial statistical analysis(map based algorithms), and others.

• Manager and Head of Retail Modeling Science Group in Mathematical Sciences. Responsible for over 100 billion in sales predictions across more than 30 customers.• Determine strategy for algorithm deployment and help drive new research to improve existing prediction engines.• Devising new machine learning approaches to solve existing problems and directing Data Science research projects with junior members. Includes Support Vector Machines, Mixed Effect Models, Fourier decomposition, etc. • Direct research to drive new products or new applications of existing products.Individual Contributor work• Coding in Java to implement prototype models into Java production environment as well as fixing mission critical bugs• Served as in-house onsite senior consultant to Walmart Sam's Club for customer insights and membership trends• Ported production imputation algorithms into Hadoop MapReduce and benchmarked results on a small cluster for proof of concept prototyping for scalable parallel processing. Also assisted port of Mixed Effects algorithms in IBM's parallelized MapReduce solver, SystemML • Investigate and improve core production Bayesian Mixed Effects estimation through MC simulations• Hold Java seminars to introduce and refine Java and OOP skills to rest of Science Team. • Created set of R libraries to automatically translate company production S plus code in R to reduce company licensing costs

• Creating Predictive models to model Property and Casualty Losses. Modeling techniques used include Generalized Linear Models, Neural networks, ARIMA time series models, Principal component regression and linear regression.Some sample projects include• Successfully demonstrating significant lift to company’s current Personal Auto models using external vendor data.• Improving Indications models by creating ARIMA and time lagged neural network models for Personal Auto accident frequency time series data.• Using Neural networks models to achieve significant lift to company’s current Personal Auto models.• Using survey data to assess personality and create new factors to improve predictive models.• Analyzing and classifying telematics data to assess risk.Other Responsibilities• Mentoring junior level employees and leading project direction• Creating new project proposals for research pipeline

Modeling, Mathematical, and Statistical Responsibilities• Single handedly created the company’s first real-time neural network fraud engine to scan incoming ACH transactions and flag high risk transactions for fraud team investigation. Delivered network which could be integrated into existing software infrastructure.• Single handedly created, implemented, and maintained first decision tree based rules used by company’s rules engine to analyze all ACH transactions and flag possible risky transactions for follow-up • Made change to existing company Trial Deposit(new account verification) distributions to save company additional 500K a year and minimize risk of fraudulent enrollment (Mathematica and SAS simulations)• Recommended and investigated analytical methods to refine fraud detection for online transactionsDatabase, Reporting, and Operations Responsibilities• Planned and created extensive geographical database(Oracle) used to obtain geographical statistics for online transactions using captured IP addresses and geo-data from third-party vendor• Create ad-hoc reports using SQL and SAS for risk assessment – Highly proficient with SQL and SAS • Single handedly created in-house ACH database and case management system to facilitate communication between Risk management and ACH to guarantee reliable company Loss Reporting to upper management.(ACCESS Database) • Monitor weekly ACH transaction data to ensure that returned item statistics are within normal ranges

• Implement and utilize SPC (Statistical Process Control) methodologies - Acquiring, storing, and monitoring process variables to identify out-of-control conditions on metrology tool using Matlab • Recommend changes to engineering team to stabilize and change process parameters based on SPC results.• Research and Perform multivariate statistical analysis of wafer process data to model and control etch processes – principal component analysis(PCA) and principal component regression(PCR) • Prediction of process gas flow rates using OES (Optical Emission Spectroscopy) signals with Partial Least Squares; • Develop Neural network models using Matlab; use in conjunction with PLS to improve process prediction Other Process Engineering Responsibilities• Develop new software tools to enhance corporate efficiency and productivity.• Interact with customers to develop etch processes using company resources.• Design and perform experiments to evaluate in-house development tools; recommend changes to hardware engineering staff based on results.