Jeremy Guinta

Jeremy has managed and analyzed complex, disparate data for a wide variety of cases, including the Fair Credit Recording Act (FCRA), privacy, cybersecurity, labor and employment, consumer protection, the False Claims Act (FCA), and general class actions. He also retains a wide variety of technical skills and tools used to perform his work. He is an expert statistical programmer and database administrator. He has also developed and delivered national training curricula for technical programming skills.  As the lead instructor for these trainings, he has successfully instructed on data management, statistical analysis, and graphics.

Jeremy also holds a position as an Adjunct Professor at Pepperdine University in Malibu. He teaches courses on advanced structured query language programming, database design and validation, and predictive analytics. He has previously held a Lecturer role at California State University, where he taught applied statistics.

Jeremy’s professional experience includes:

Labor and Employment Experience

• Replicated a major hotel chain’s payroll database and utilized all available data to compute meal and rest break violations and penalties in accordance with state law. Analyzed payroll, timekeeping, and human resources data from multiple disparate systems into an electronic database ready for analysis.
• Conducted a multi-state, cross-jurisdictional investigation of a major United States retailer’s point-of-sale, inventory, and training system data, and timekeeping system data, to determine unlawful off-the-clock hours worked by employees.
• Analyzed human resources data related to a reduction in force by a major auto parts manufacturer that the plaintiffs contended was discriminatory.
• Prepared analysis to assess the potential liability of non-exempt employees in a class action labor dispute concerning gender pay discrimination. Developed a regression model that estimated the potential disparate impact on females based on a large oil and gas company’s payroll practices. Using a log-linear model, Jeremy estimated the percentage difference between men’s and women’s average annual wages. He determined key features that could be used to explain the alleged discriminatory effects.

Consumer Class Actions

• Reviewed, processed, and analyzed dialer and leads records in excess of 500 million observations from a national insurance agency facing Telecommunication Consumer Protection Act violations. Rebutted plaintiffs’ expert’s opinions regarding the number of calls made.
• Analyzed Binary Large Object (BLOB) files and traced information from BLOB to processed data to output report on behalf of a company facing FCRA litigation regarding inconsistencies across key pieces of information that led to the alleged denial of credit. Jeremy analyzed the data and documented the processing and how information was gathered, stored, and aggregated for a consumer.
• Implemented an ML algorithm that used natural language processing to detect Protected Health Information (PHI) across the company’s systems. This process derived features from various metadata, NoSQL objects, and rows and columns of structured data. These features were then used as predictors in a larger model to pre-tag observations as containing PHI or not containing PHI.

Regulatory and FCA

• Testified in a matter that involved FCA allegations against a provider of mental health services in Massachusetts facing FCA allegations. Jeremy reviewed the underlying data, the plaintiff’s claims matching process, and the plaintiff’s penalty calculations. Jeremy prepared an expert report and gave deposition testimony regarding the plaintiff’s methodology. He also opined on the plaintiff’s lack of quality review and data validation of the underlying sources.
• Analyzed sales data, ordering data, customer review data, and shipping data, and used customized tools to gather third-party reviews from online sales portals on behalf of a company that was facing a large FTC investigation into alleged violations of the mail-order rule and allegations of negative review suppression for reviews on the company’s website.

ML, Statistics, and Econometrics

• ML: Designed, trained, and tested a supervised ML algorithm to assist with the review of names across multiple data sources and to determine whether a name in one data source matched a name in different data sources.
• Sampling: Devised a statistically valid sampling methodology to select paper time sheets for data entry and analysis. Jeremy used various tools and the known population size of documents to draw the sample. The sampling procedure reduced the number of documents from hundreds of thousands to a few thousand but achieved 95% confidence with plus or minus a 5% margin of error.
• Econometrics: Pioneered various approaches that measure damages under evolving UK case law. These approaches included using Last-In, First Out, First-In, First-Out, and Weighted Average accounting methods to account for purchase/sale pairings of share transactions. Jeremy developed and implemented econometric event studies to measure the alleged inflation of the share price based on information being withheld from the market due to the alleged fraud.