• Planned and completed a major, 15-month migration project from Google Analytics to an in-house server-side eventing system, increasing visibility into user behavior by ~15% and reducing system complexity by two orders of magnitude, achieving $160,000 in annualized savings.• Mapped out development of an AI-based conversational agent from a simple multi-index, single-retrieval solution to an advanced agent-based multi-index, multi-retrieval and led productionalization of first two… Show more • Planned and completed a major, 15-month migration project from Google Analytics to an in-house server-side eventing system, increasing visibility into user behavior by ~15% and reducing system complexity by two orders of magnitude, achieving $160,000 in annualized savings.• Mapped out development of an AI-based conversational agent from a simple multi-index, single-retrieval solution to an advanced agent-based multi-index, multi-retrieval and led productionalization of first two iterations of the product. Show less

Guided an 8-person team of Data Scientists, Analytics Engineers and Business Analysts providing business stakeholders across all of RealSelf’s teams with intuitively interpretable data products and actionable insights enabling and promoting data-driven decision making.• Provided guidance on a revitalized Business Review format directly linking individual team KPIs to company-level KPIs that ladder up into OKRs, ensuring clearly defined ownership of metrics and improved… Show more Guided an 8-person team of Data Scientists, Analytics Engineers and Business Analysts providing business stakeholders across all of RealSelf’s teams with intuitively interpretable data products and actionable insights enabling and promoting data-driven decision making.• Provided guidance on a revitalized Business Review format directly linking individual team KPIs to company-level KPIs that ladder up into OKRs, ensuring clearly defined ownership of metrics and improved accountability.• Led strategy to reduce costs and optimize BigQuery usage for a 65% direct cost reduction within 2 quarters, translating to $240,000 in annual savings.• Directed the migration of key reporting pieces from existing BI platform (Looker) to a no-cost-per-query tool with improved UX and functionality (Falkon), cutting license seats by 40% & saving $20,000 per year.• Initiated development of collaborative and content-based filtering-based recommendation engine providing personalized recommendations across all content types, leveraging GCP’s Retail API and underutilized spend commitments. Show less

Led a 4-person team of Data Scientists and Analytics Engineers maintaining a SQL-centric model stack and supporting internal and external stakeholders with recommendations, forecasts and advanced analytics.• Oversaw the enhancement of customer profiling and email content recommendation systems, serving recommendations to ~5M users per week and driving 6% of overall site traffic.• Supervised improvements to dynamic share-of-voice system governing the company’s main… Show more Led a 4-person team of Data Scientists and Analytics Engineers maintaining a SQL-centric model stack and supporting internal and external stakeholders with recommendations, forecasts and advanced analytics.• Oversaw the enhancement of customer profiling and email content recommendation systems, serving recommendations to ~5M users per week and driving 6% of overall site traffic.• Supervised improvements to dynamic share-of-voice system governing the company’s main impressions-based advertising product driving $2M in MRR.• Successfully managed implementation of a lead routing system for key accounts across multiple ad products, significantly improving individual customer ROI and reducing churn by ~7% Show less

• Created a technical interview preparation program, developed & revised frameworks for deep dives and business cases, and delivered to 200+ Data Science and Health Data Fellows across the entire company, helping to lift overall placement rates by 80% relative to previous year and 240% relative to previous session.• Supervised transition of Boston Health Data program to fully remote model in response to COVID-19 pandemic, creating company’s most successful program (out of 19) and… Show more • Created a technical interview preparation program, developed & revised frameworks for deep dives and business cases, and delivered to 200+ Data Science and Health Data Fellows across the entire company, helping to lift overall placement rates by 80% relative to previous year and 240% relative to previous session.• Supervised transition of Boston Health Data program to fully remote model in response to COVID-19 pandemic, creating company’s most successful program (out of 19) and improving placement rates by 107%. Show less

• Mentored and advised 100+ junior data scientists on product-centric data science projects involving various machine learning techniques across multiple technical domains (regression, classification, times series/forecasting, natural language processing, and computer vision).• Developed and maintained relationships with 60+ partner companies of varying sizes (small startups to large enterprises), working with hiring managers to target the right talent and build highly cross-functional… Show more • Mentored and advised 100+ junior data scientists on product-centric data science projects involving various machine learning techniques across multiple technical domains (regression, classification, times series/forecasting, natural language processing, and computer vision).• Developed and maintained relationships with 60+ partner companies of varying sizes (small startups to large enterprises), working with hiring managers to target the right talent and build highly cross-functional teams.• Managed end-to-end admissions process from application to onboarding for 4 Fellowship cohorts, interviewing 150+ candidates for highly selective data science and health data programs. Show less

• Developed a forecasting app in Shiny that strategically informs government planning by predicting counties at high risk for increased overdose mortality rates to target for expansion of medication-assisted treatment (MAT) service systems.• Integrated, cleaned, and explored several disparate data sets on mortality, opioid prescribing rates, and sociodemographic features from Centers for Disease Control and Prevention, US Census Bureau, and Substance Abuse and Mental Health Services… Show more • Developed a forecasting app in Shiny that strategically informs government planning by predicting counties at high risk for increased overdose mortality rates to target for expansion of medication-assisted treatment (MAT) service systems.• Integrated, cleaned, and explored several disparate data sets on mortality, opioid prescribing rates, and sociodemographic features from Centers for Disease Control and Prevention, US Census Bureau, and Substance Abuse and Mental Health Services Administration using dplyr and ggplot2.• Built additive regression model based on historical trend information in combination with several additional regressors to predict mortality rates at the county-level using Facebook Prophet, reducing the average MSE of prediction by 250% relative to a simple linear model based only on year and prescribing rates. Show less

• Analyzed RNA-sequencing data from patients receiving direct-acting antiviral treatment for chronic Hepatitis C virus (HCV) infection using linear models, clustering, and dimensionality reduction to detect patterns of incomplete immune recovery after successful treatment.• Utilized gene-correlation network analysis to survey RNA-sequencing data from rare T cell populations to detect disease outcome-specific gene expression patterns and identify high-value gene regulatory candidates for… Show more • Analyzed RNA-sequencing data from patients receiving direct-acting antiviral treatment for chronic Hepatitis C virus (HCV) infection using linear models, clustering, and dimensionality reduction to detect patterns of incomplete immune recovery after successful treatment.• Utilized gene-correlation network analysis to survey RNA-sequencing data from rare T cell populations to detect disease outcome-specific gene expression patterns and identify high-value gene regulatory candidates for clinical intervention. Show less

• Investigated T cell transcriptional profiles by applying a combination of supervised learning, dimensionality reduction, clustering, and weighted correlation network analysis techniques to identify distinct gene regulatory processes associated with chronic viral infection.• Developed a complex analytical pipeline linking gene expression patterns to clinical predictors of disease outcome by integrating data from multiple experimental platforms with clinical metadata and information… Show more • Investigated T cell transcriptional profiles by applying a combination of supervised learning, dimensionality reduction, clustering, and weighted correlation network analysis techniques to identify distinct gene regulatory processes associated with chronic viral infection.• Developed a complex analytical pipeline linking gene expression patterns to clinical predictors of disease outcome by integrating data from multiple experimental platforms with clinical metadata and information from external databases, generating new functional insights gained from established assays. Show less