Professor of Record for the Data Science MS program at the Seidenberg School of Computer Science and Information Systems. Semester-long 3-credit required class to prepare students for their Analytics Capstone Project and post-graduate career.Course entitled “Practical Data Science”. Developed the first-ever required class, for this MS program, that prepares graduate students for their first full-time Data Scientist position in a corporate setting.

Performed data science projects in the Healthcare Analytics space.

Recency, Frequency, & Engagement (RFE) measured the health of ESPN+ subscribers to create target audiences for tailored ad campaigns aimed at preventing churn. RFE features quantified when users streamed content, read articles, or engaged in fantasy sports on the ESPN website/app. Using these features, a K-means algorithm clustered users into groups that revealed distinct trends in content consumption, which were associated with specific churn levels. These findings enabled the creation of segmentation logic. Every day in production, this logic was used to automatically segment users into healthy, unhealthy, or do not disturb groups, based on their last 28-90 days of activity.Participated in a 72-hour Datathon and won the Highest Value award for delivering a working POC. Multi-class text classification models assigned the work type field in 3K incomplete Jira tickets based on each ticket’s text fields. All text pre-processing and modeling code were written in PySpark to maximize speed and efficiency in Databricks. Subsequently selected as team leader to refine and build this process in production.Built the first-ever POC that used ESPN clickstream data for ESPN+ acquisition modeling. Expected to have a causal lift in ESPN+ subscriptions totaling $7M+, when deployed. Built a Spark-based pipeline that wrangled and modeled trillions of records. Evaluated results from logistic regression, random forest, and gradient boosted tree (GBT) models; GBT was selected. Model performance significantly improved after handling class imbalance via downsampling in training; predicted probabilities were calibrated to correct for bias introduced in the posterior. To optimize hyperparameter tuning, a Bayesian approach, via Hyperopt, was used over traditional grid search. Final user-level predictions were bucketed into likelihood categories to define target audiences for tailored ads and feature importance was analyzed to customize display ad content for experiments.

Survival Analysis: Analysis used by Finance to determine existing content’s value and assess new comparable content to acquire. Tracked survival rates of ESPN+ users, based on their content consumption by time period. Built a Dash app so end users could select content watched and time period to run the model in real time and generate week over week survival curves.Project Management: Led data science projects from inception to model deployment. United stakeholders from data science & engineering, product management, and marketing around common goals to build models that solved business problems. Proficient in communicating project status, timeline, expectations, and findings, while executing all data science tasks.App Development: UFC Dash app was a key tool for leadership to monitor near real-time acquisitions. Known as a “go to” Dash app developer in the broader DATA org. Successfully designed, built, and deployed the first Disney Streaming Dash apps in production to provide management with interactive user-friendly apps that showcased different cuts of real-time data to enable business decisions.Leadership: Provided technical guidance to Data Analysts to develop their knowledge of statistical methods, Git, Databricks, and Python.

Awarded Workday badges for project management and delivering tangible & impactful outcomes.Sentiment Modeling: Built a sentiment analysis model in SageMaker to extract structurally negative reviews based on a vocabulary related to 5 types of insurance liability risks. Generated numeric risk scores for each liability risk as well as an overall risk score for each business using negative reviews. Results were integrated into a user interface for underwriting a policy. HR Analytics: Analysis demonstrated that an organization can leverage anonymous employee survey data to create better employee experiences. Analyzed raw employee experience survey data to provide insights into how different groups of employees viewed diversity, inclusion, engagement, and opportunities within a firm. Developed R code that explored these trends using multinomial logistic, lasso, and ordinary least squares regression with cross validation, as well as various correlation tests.Time Series Analysis: Weekly predictions would have alerted the supply chain team to future stockout or overstock situations in advance, so inventory levels could be adjusted to avoid inventory misses. Built an ARIMA model in Azure ML Studio that predicted weekly inventory levels for each item at each retail store for the next 16-weeks. Data Strategy: Recommended strategies to improve data consistency and quality to enable marketing to target specific audiences for ad campaigns to increase reach and revenue. Analyzed how digital content data were tagged/categorized in Snowflake to showcase inconsistent tagging practices across multiple brands.

Utilized text mining to automate the Quarterly Performance Report, reducing preparation and monitoring time by 90%.Obtained data via API calls to analyze, interpret, and translate data mining results into data visualizations; recommended statistical modeling approaches to understand Superstorm Sandy Relief Program participant behavior.Prepared and delivered weekly R programming and data retrieval tutorials to develop the team’s coding proficiency.

Accomplishments: Promoted in 11 months, ranked as a top tier performer within Senior Auditor peer group, was known as the “go to” auditor who supported senior management, and was sought after by managers to lead audits. Well versed in synthesizing multiple stakeholder inputs to deliver a final polished product. Led a 3,000-hour end-to-end audit of Global Commercial Mortgage Lending: Managed 5 internal audit and 5 client teams to identify and remediate control weaknesses that may cause regulatory, financial, and/or reputational harm to AIG.Spearheaded a 1,500-hour Investment Compliance audit to evaluate and remediate control gaps in the monitoring process of insider trading. Analyzed raw username data, from multiple systems, to create a database of employee names that highlighted significant gaps in compliance with the SEC Investment Company Act of 1940; finding was ranked as the #1 audit issue.Managed comprehensive audits by driving project plans to uncover control gaps. Collaborated with the business to create and implement action plans that resolve audit findings and improve processes.Trained new team members to complete audits i.e., work paper documentation, sampling techniques, and issue identification, remediation, and verification.

Analyzed various assets tied to emerging market economies to propose investment recommendations based on hedging strategies informed by Modified Sharpe Ratio and regression analyses. Determined valuation of assets based on political risk, volatility, weekly price, and net asset value. Created regression models to understand how benchmarks impact prices, including gold prices, emerging market equity indices, and interests rates.Completed audit work for CCAR, Life Settlements, Residential Whole Loans, Investment Management Agreement (IMA) limits, Private Placements, Affordable Housing, Real Estate Mortgage Investment Conduit (REMIC), and Derivative Operations. Highlights from audit work include:Examined CCAR stress scenarios, reinvestment calculations, assumptions, inputs, rationale, and results related to the Chief Investment Office in order to verify that previously raised regulatory issues were properly remediated by the business. Verified that appropriate structured products price modifiers were used by the Private Placement Group. Aggregated and analyzed monthly loan delinquency data to determine if proper loss mitigation actions were taken.Re-performed daily rate sheet pricing calculations to verify that EUTs & Rate Sheet Builders were designed and operating effectively to mitigate market risk.Reconciled discrepancies between various loan sub-servicer trial balance reports to verify that Total Portfolio Balance Verification was performed at month end. Ensured discrepancies were properly investigated, remediated, and aged to mitigate operational risk (SOX control).Verified SFTP (SSH File Transfer Protocol), Life Settlements Database, and ShareDrive user access was appropriate based on employee roles and responsibilities to protect Personally Identifiable Information.

Supported CCAR with variance analysis to assess the accuracy of AIG’s financial forecasts vs. actual and evaluated output variances between stress models to enhance controls. Also examined stress scenarios from the Federal Reserve and AIG to identify/predict the impact of adverse economic shocks from an insurance and bank holding perspective.Evaluated Anti-Money Laundering, Bank Secrecy Act, and Economic Sanctions Program policies for the Federal Saving Bank to verify completeness, adequacy, and compliance with AIG’s global policies and regulatory requirements. Assessed the OFAC screening control environment for incoming and outgoing wire transfers.Completed tasks for PwC Reliance under budget and ahead of schedule. For derivatives and hedging instruments, completed a 10-Q Disclosure Tie-Out and tied out variances between journal entries and source data to understand the impact on Q2.In addition to regular audit-related responsibilities, supervised 14 interns in New York and Houston to develop the company’s first internal audit diversity video. Also completed training on derivative accounting, high-grade/high-yield financial products, and dealing with difficult clients.

Oversaw a time-sensitive, unit-wide audit process to identify and document material inadequacies of financial statements against SEC and FINRA standards. Analyzed key variances on Exception Disclosure Reports to identify credit, market, and liquidity risk exposures.Highlights of this position include collaboration with the Executive Vice President and Examination Director to design and implement an internal control. Also formalized unit-wide Profit & Loss tie out analysis template and annual audit review procedures.

Recognized as the top ranking Intern of Internal Audit Division as scored on final intern assessments. Supported senior management in assessing the control environment by populating Risk Assessment Matrices, documenting issues, and verifying audit issue action plans.

Conducted approximately 400 cold calls per day in a fast paced financial environment to recommend strong prospective clients to wealth managers.

Helped recruit American fashion designers for the 2013 Peru Moda Trade Show and supervised interns by leading group presentations.