Harrison Wallace Email & Phone Number

Chicago, Il, Us

Chicago, Il, Us

▪ Designed a gradient boosting machine (GBM) model to incorporate static variables about each particular NFT, dynamic real time sales information, and macro level market wide information to predict sales price.▪ Incorporated over 13 million prior sales over a 12 month period to predict future sales.▪ Productionized model to pull latest data for batch level scoring every hour to be displayed on the website; automatic model refits on a rolling latest 1 year of data occur on a weekly basis.▪ Designed a suite of analytical visualizations produced during automatic model refits to enable future model owners to be confident in the continued reliability of the model’s predictions, with special attention paid on correctly identifying the outlier most valuable NFTs on the market.

Mclean, Va, Us

▪ MRO High-Level Process: Collaborated with business and DS team associates to measure quantitative data and qualitative risks, and model integrity.▪ Reviewed business objectives, model design, data design and infrastructure, coding, model fitting best practices, and evaluated results; providing guidance and direction to business analysts and data scientists responsible for submitting their work on model builds.▪ Took submitted models and built challenger models to evaluate potential weaknesses, observe how models fit across various segments of interest; created hypotheses for improvements; used regression (GLM, GAM) and GBM models. ▪ Created and represented PowerPoint presentations and reports demonstrating metrics, analytics, predictive outcomes, recommendations and solutions; deliver clear and articulate presentations at team meetings.▪ Submitted model validation reports to internal auditors and external regulators, e.g., Federal Reserve, OCC.▪ Partnered with engineering, product, and design teams from across the company to validate the development and deployment of models; collectively built, tested, and publicized tools and systems to optimize the model risk management process.▪ Utilized predictive analysis and model risk evaluation to effectively support the company’s vision, strategies and innovation, i.e., to bring ingenuity, simplicity and change to banking that allows people to live a better lifestyle.▪ Expert knowledge and experience with internal risk models rank ordering probability of default, valuation models measuring net NPV over the lifetime of an account, and FICO / VantageScore publicly available credit risk models.▪ Ensured software followed best practices, e.g. proper version control and unit testing, replicable results; second line validators were able to properly evaluate first line model developer’s technical infrastructure.

Mclean, Va, Us

▪ Team member responsible for rebuild of a valuation model built in SAS that covered multiple portfolios▪ Moved to the Kohl’s team, one of the largest credit cards in Partnerships, tens of millions of open cards; managed multiple models; frequently evaluated appropriateness for task and inaccuracies; determined whether a model rebuild was necessary.▪ Worked closely with business analysts to conduct data analysis and craft credit policies through use of valuation models; guided the insight and integration between data scientists and business teams.▪ Designed and launched a Foundational Test to provide credit limit increases to people in a randomized way regardless of whether results will be profitable as part of a test market strategy to be used to more accurately measure probability in future decisions.

Mclean, Va, Us

▪ Conducted queries in Teradata SQL databases containing millions of rows and complex table to table relations; required management of data in SQL and SAS, with heavy emphasis in automation. ▪ Supported business analysts in complex data analysis and interpretation of coding.

Mclean, Va, Us