Adam D'Amico Email & Phone Number

Richmond, VA, US

Richmond, Virginia, United States

• Predicted inbound calls using cohort/grouped data and a limited feature set at the campaign level using OLS (99.2% accuracy), then at the cell level (~550) using XGB (87.4% accuracy) to optimize quantities and.
• Developed data-rich dashboards and streamlined reporting utilizing Tableau, Python, and Jupyterlab
• Aligned the DM matrix and SQL code to select stratified random samples and A/B test quantities at the cell level - reduced campaign processing time by 15% and increased response rates by 3.1%
• Built ML pipelines to score models across multiple markets and identified consumers with the greatest likelihood to convert to customers
• Aggregated data from multiple public APIs (Census, BLS, CEX, etc.) to create features for the entire Data Science team - used in analyses and model building to explain and control exogenous factors
• Developed a PDF Generator complete with graphics and a Decision Tree model to contrast clients’ customers vs consumers and analyze key marketing messages to resonate with target segments

Richmond, Virginia, United States

• Built model to predict days between response and application using Acxiom data, leading tosmoother lead distributions for sales and more accurate application forecasting for Marketing
• Aggregated data from multiple public APIs (Census, BLS, etc.) to create features for the entireData Science team to include and test in model building and analyses
• Developed system to productionalize models, from writing the SQL, pulling the data, runningthe models, scoring the leads, and ultimately posting scores to our Salesforce.com REST API
• Built functionality to assess profitability at a policy level, starting and stopping at any monthin time, for any cross-section of the entire portfolio
• Utilized pdfminer and pdfplumber to extract text from all product versions of applicationsand build a database to test the efficacy of questions and wording/phrasing changes over time
• Developed a scorecard analysis process to test model efficacy with appropriate statisticaltests, graphs, and distributions for all models in production

• Software Engineer: Actuarial IT Support
• Developed Proof of Concept dashboard for a complex job involving actuarial processes,reducing time to diagnose errors by 14%
• Built process to funnel data and outcome metadata into Splunk for reporting and analyticsWeb Development
• Made code changes to >1,000 ADA issues in order to have accessibility options for customers
• Developed new pages around MayoClinic’s Live+Well API
• Created and executed unit tests, bug fixes, and sat in on deployments