Isaac Wright, Phd Email & Phone Number

Boston, MA, US

• Collaborate with actuarial teams to develop and enhance reserving models using advanced statistical and machine learning techniques.
• Analyze and interpret large datasets to identify patterns, trends, and insights that inform the reserving process.
• Build predictive models to estimate and forecast claims and loss reserves, incorporating various factors such as policy information, historical data, and external variables.
• Apply data science methodologies to optimize reserving processes, including data preprocessing, feature engineering, model selection, and validation.
• Conduct thorough research and stay updated on emerging trends, technologies, and best practices in data science and actuarial science fields.
• Present findings and recommendations to stakeholders, including actuaries, underwriters, and senior management, in a clear and concise manner.

University Park, PA, US

• Collaborate with lead faculty on a $250,000 grant to develop a model for predicting messages that traders send to exchanges during periods where there is a shock to liquidity. This model will help regulators develop.
• Leverage computing skills in R to apply machine learning (ML) techniques to classify different trader types in financial markets based on limit order books, including high frequency traders, liquidity suppliers, market.
• Implement a Bayesian mixture model setup to identify clusters of similar limit order book.

University Park, PA, US

• This MS Thesis focuses on the application of Machine Learning techniques to financial data. The abstract is below.
• Build a classifier to predict the direction of stock market returns using past prices and volume as predictors.
• Implemented techniques such as Random Forest, Boosting, and Quadratic Discriminant Analysis to evaluate the first level of the EMH called the Weak EMH.
• Wrote R code for parallel computing to increase computational speed and significantly reduced run time.
• Previous research in the area focused heavily on linear models and used their failed models as evidence of the EMH. This thesis takes an alternative approach an instead uses non-linear models since there is no reason.

University Park, PA, US

• Carry out statistical analysis and work with the client to interpret the results and how they impact their research.
• Report findings back to a client, whether it is using a formal report, a client presentation or simply a conversation over the phone.
• Establish ongoing client relationships to encourage longer-term working partnerships.

Boston, MA, US

• Improved overall accuracy of predictive models by 12%.
• Designed and implemented novel Machine Learning methodology to boost predictive performance for Catastrophe severity models.
• Gathered, organized, and analyzed large data sets to create highly predictive features for machine learning models.
• Revised and revamped previous severity models to take advantage of newer data and previously unused data structures.
• Communicated business impact of new methodology to leadership team through presentation deck.

Boston, MA, US

• Met with various business units to scope a project that would provide the business with significant cost savings.
• Built a modeling data set by implementing data wrangling techniques in R to combine multiple data sets into one.
• Leveraged my statistical background to train and evaluate several machine learning (ML) models' usefulness in providing data-driven guidance to underwriters.
• Produced a cost-benefit analysis (CBA) comparing the ML models to current guidance to show the models would provide an estimated $1.1 Million in annual savings (~49% cost reduction).
• Created a presentation deck to share with leadership to communicate the significant cost-savings a data-driven approach would provide to the business.

Savannah, Georgia, US

Savannah, Georgia, US

Project Title: The Fascinating Gamma FunctionThis paper will discuss properties of the Gamma Function and provide proofs for each property.

Ann Arbor, Michigan, US

• Derived and coded a likelihood function for bivariate censored data using optimization tools in R software.
• Coded simulations to assess the accuracy of the likelihood function for censored data.
• Presented research on censored data solutions at the SROP 2016 Symposium.

Columbus, Ohio, US

• Assisted research professors in Mathematics and Statistics with statistical analysis of data.
• Developed computer programs in R for sorting and interpreting numerical data files.
• Presented research findings at the SROP Symposium.

Savannah, Georgia, US

• Advice peers academically and socially.
• Tutor for Mathematics and Computer Science classes.
• Coordinated and hosted final exam workshops for students in mathematics and statistics courses

University Park, PA, US

• Eight-week intensive summer research project working in Penn State’s Fluid Mechanics lab.
• Conducted multiple experiments and analyzed data using Mathematica programming languages
• Presented research on decay rates of shallow-water waves at the SROP Symposium.

Nashville, TN, US

• Reinforce learning of mathematics/statistics material or skills initially introduced by the teacher in small groups and with individual students.
• Assist the teacher in developing strategies for reinforcing mathematics/statistics material or skills based on student interests and abilities.
• Coordinates outreach activities for students to gain exposure to careers in the STEM fields.

Doctor Of Philosophy - Phd, Statistics

Master Of Science - Ms, Statistics

Bachelor Of Science (Bs), Mathematics