Ed Chow Email & Phone Number

Sr Computer Scientist, Artificial Intelligence Center

Directed the development and delivery of the predictive analytic models underlying FICO’s Customer Lifecycle products in the Originations, Account Management, and Collections/Recovery domains in the banking and financial industries. Supervised teams of analytic scientists, from entry level to Principal Scientist, overseeing development and implementation of score cards and transaction models; project delivery; process improvement; client support; and research in data mining and statistical modeling. Represented FICO analytics in critical client-facing meetings and high-profile projects. Interfaced with internal product management, software development, product delivery, operations, analytic consulting, and sales and support teams. Provided analytic guidance to product development teams and strategic corporate initiatives.

Manager, Advanced Technology Solutions Group and Technology Transfer Group. Performed research into natural language understanding and semantic relationships within unstructured data. Developed software tools for extraction of semantic representations from large text corpora. Prototyped graphical user interfaces for visualizing information-theoretic relationships, for demonstrating named-entity recognition and disambiguation, and for visualizing high-dimensional data. Used machine learning tools and novel text mining algorithms to create predictive models and self-organizing maps from unstructured text data. Conducted research in the Pairwise Co-occurrence Consistency algorithm for measuring the strength of links in large amounts of retail transaction data, with application to recommendation engines.

Responsible for development of the predictive matching engine behind the Lens suite of résumé-processing products. Created neural network models to assess the degree of match between résumés and postings. Built statistical models to predict employee longevity. Assisted in development of the Aperture résumé processing system and its web interfaces. Implemented text classification algorithms (Naïve Bayes, Hierarchical mixture models) for categorization of job postings into the O*NET occupational hierarchy and developed a confidence model that returns the probability of a correct classification.

Continued to improve, extend and innovate the AREAS suite of valuation models. Revamped the model development process using a new modularized environment, and developed a new Index model and Quality rating. Expanded the national coverage from 100 to almost 300 counties. Enhanced the product with improvements in accuracy of models, efficiency of data processing, automation of monthly update and deployment, and performance of databases.

Created predictive neural network models for the AREAS automated real estate valuation product. Prototyped, experimented with, and implemented new techniques for improving the accuracy of model predictions. Streamlined and automated the process of collecting and inspecting data, of building models, and of evaluating model performance. Provided consultation and technical assistance to clients.

Assistant Professor, Queen’s University, Department of Mathematics and Statistics, Kingston, Ontario, Canada. Performed research in theoretical and computational statistics. Taught courses from the first year undergraduate to Ph.D. level, from elementary service courses to research seminars. Supervised two Ph.D. students, two M.Sc. students, and multiple applied undergraduate projects. Provided extensive consulting help on advanced statistics, programming, and statistical software to students, faculty and staff. Co-organized the Queen's University Statistics colloquium; chaired the Institute of Mathematical Statistics New Researchers' Committee (1994-1996); and organized the 2nd North American Meeting of New Researchers in Probability and Statistics (July 1995).

Ph.D., Statistics

B.Sc. (Honors), Mathematics And Computer Science