Paul Torres Email & Phone Number

San Francisco, California, Us

New York, Us

- Redesigned analytics data layer resulting in a 60% decrease in run time and a 30% decrease in daily AWS costs- Led analysis of new onboarding system, which required: * custom AB testing pipelines to account for restrictive front-end requirements * built clear and concise visualizations for consecutive tests where variant groups often interacted * advised engineering and product partners on the quality data best practices- Created a new attribution system that used last touch attribution and made allowances for new privacy restrictions for iPhone users that enabled the company to allocate marketing resources efficiently- Conducted deep data analysis on user cohorts, finding that depositing users have a 7x longer lifespan on the platform than those who only completed their account, leading to a shift in marketing strategy.

New York, Us

- Oversee acquisition data science needs — including data infrastructure, stakeholder-facing dashboards, and analysis- Helped rewrite attribution model to use last touch instead of waterfall touch points- Drafted, designed, and executed the release of a self-service curriculum for data education company-wide

New York, Us

- Led data science analytics for product marketing to drive and report activation and retention metrics- Brainstormed and led the investigation into decreased messaging numbers — increasing activation rates as a result- Launched investigations that led to rewriting of security rules and gave 20k+ users access to the platform- Designed and completed an overhaul of messaging data to help speed up the evaluation of marketing campaigns- Conducted weekly data tool lessons for product managers to help learn self-service data needs- Oversaw data management for crypto launch for Robo Invest on the platform and delivered causal impact analysis

- Developing an application that reads a PDF and auto-fills out a form to be used transforming medical reports into tabular data- Created a clustering algorithm that identified food deserts in New York City

• Gather census and ACS data from 2000, 2008, 2010, and 2012 that includes income, housing, and ethnic demographic data• Calculated percentage and percent changes for the representative data• Ran clustering algorithms to find which subset of features explained the most variance and which clustering algorithm created the tightest clusters with the highest silhouette score• Visualized clusters with Tableau software to compare with criteria set by academic papersTech Stack:• Pandas• SQL• PostgreSQL• SciKit-Learn• KMeans Clustering• Tableaugithub.com/ptorres001

New York, New York, Us

Complete 15-Week of intensive training to develop a theoretical and practical understanding of programming, statistical analysis, machine learning, and deep learning.Select projects include ...0. MLB Salary Predictor (Linear Regression, RFE, VIF, Data Collection, Feature Engineering)1. Pneumonia Image Classification (CNN, PCA / Eigenimages, TensorFlow, Keras)2. NYS Health Insurance Classification (Lasso Regression, kNN, Decision Tree, Random Forest Classifier, XGBoost, Voting Classifier, Feature Engineering, Pipeline)3. Beijing Pollution Level Predictor (SARIMA, Facebook Prophet, LSTM, Plotly)4. Neighborhood Gentrification Clustering Identification (KMeans, PCA, KNN, Hierarchical Agglomerative Clustering)5. Credit Card Default Prediction (Logistic Regression, kNN, Random Forest, XGBoost, Feature Engineering)

• Worked with 2010 Census Data in an effort to correctly classify whether a person was covered by health insurance in New York State• Preprocessed data to differentiate between continuous and categorical data• Tested different models during cross validation in order to find the best model to predict the coverage standing• Deployed best model on data from Alabama to inspect how it would perform in a place where the health care policies were very differentTech Stack:• Python• XGBoost• Tableau• Light Gradient Boosting Machine• Random Forest Classifier• Logistic Regressiongithub.com/ptorres001

• Used baseball statistics from 2000-2019 to predict what a rookie's salary would be after arbitration• Combined several datasets that included batting statistics, pitching statistics, vitals, and awards• Engineered advanced statistics that would be helpful to predict salary• Ran different linear regression models with feature selection attributes that would eliminate colinear features and those that did not contribute to explaining variance• Future plans include using metrics used by scouts like pitch velocity, run speed, etcTech Stack:• Pandas• SQL• PostgreSQL• SciKit-Learn• Linear Regression• Lasso CV• Gradient Boosting Regressor• Dummy Regressor• Tableaugithub.com/ptorres001

• Created time series for meteorological events like high-pressure systems• Work included data from NOAA in order to model weather system behaviors.• Univariate time series to determine geopotential height during weather phenomena • Worked on a research team that required communication and collaboration. • Took the lead on data cleaning and initial visualizations for Exploratory Data AnalysisTech Stack:• Matlab• Machine Learning• Time Series

Organized independent research on community level issues that often included local groups. Partnered with local health and education groups to better learn about the needs of their constituents. Created a health initiative in response to the community input that would better educate and source healthy foods.

I worked alongside professors in the classroom, helping demonstrate problems and solutions. Outside of the classroom I ran my own recitation sessions in which I instructed students on ways to view and solve problems.

Located sources and summarized notes for historical research with lead investigator. Analyzed and interpreted research for clear summary in order to use information for main writing project and my own paper that was presented at conferences.

Data Science Immersive Program, Data Science

Bachelor Of Science - Bs, Physics

Associate Of Science - As, Liberal Arts And Sciences, General Studies And Humanities