CEO2019 - PRESENTCedar Labs
Technologies: Python, AWS, Node.js, Serverless
- Led data science consulting practice, focusing on data strategy and growth-oriented data services.
- Served as fractional chief data officer for early-stage startups: set strategy, built capabilities and products, hired talent.
- Supported user acquisition, forecasting, resource allocation, product development; specialized in fintech and SaaS companies.
- Offered the following services; strategy, governance, hiring data teams, architecture, predictive analytics, time series analysis, regression, classification, recommendation systems, machine learning, natural language processing, ETL, and pipelines.
Lead Data Scientist2018 - 2019Foundry.ai
Technologies: Python, Scikit-learn, TensorFlow, Docker, AWS, Serverless, Node.js, Elasticsearch
- Served as general manager of hud.ai (the second oldest portfolio company within the Foundry.ai startup studio): led the product roadmap, data science, business development, and customer success.
- Built dozens of new feature sets/models across full stack, coordinating in-house and offshore developer teams and doubling sales. Executed pivot in product strategy, built new product and go-to-market strategy.
- Built NLP models with TensorFlow and Scikit-learn, deployed using Docker, Serverless/Lambda, and scaled with Node and elasticsearch.
- Developed new ETL process and built data pipelines from scratch for large text datasets, using Python, Selenium, AWS Lambda + Layers, and both Postgres and a data lake.
- Consulted for telecom company and built advanced predictive models for buying propensity, customer churn, and marketing campaign effectiveness. Built large competitive intelligence system using cloud-based crawlers, which along with first-party data were used to train and update predictive models.
Co-founder, Head of Product2016 - 2018CitySense Technologies
- Provided SaaS analytics services to water utilities to reduce lost revenue.
- Built most popular product, which forecasted resource usage and identified deviations in order to pinpoint equipment degradation.
- Built an MVP and beta app, launched in three pilot cities.
- Won third place at the Penn Wharton Startup Competition ($35,000), Summer Venture Award ($10,000), and Innovation Fund ($1,500).
Senior Product Manager (MBA Intern), US Marketplace2017 - 2017Amazon
Technologies: Scala, SQL
- Designed and built machine learning models to target grocery and CPG items to fast-growing Amazon shopper segments.
- Led partner teams around the world to implement and refine model, scaled to three international markets.
- Led process by coordinating technical and non-technical teams during scoping, development, and planning for roll-out.
Product Manager (Earlier roles: business manager, senior business analyst)2013 - 2016Capital One
Technologies: SAS, SQL, Python, Excel/VBA
- Led acquisitions of high-spending consumers across the company’s flagship products, Venture and Quicksilver.
- Created an innovative model connecting the effects of offline and online marketing. Obtained a $10 million testing budget to conduct full analysis. Led a small team in the analysis of results.
- Delivered learnings that drive millions in annual value to the CEO and Board. Developed product strategy for market-leading cash rewards card, Quicksilver, growing new customers through digital channels by 40% annually.
- Designed and launched multi-million-dollar campaigns to acquire customers and build brand value in high-impact markets, coordinating a half-dozen internal and external teams. Promoted < nine months after joining the team.
- Oversaw credit card loss forecasting using advanced financial models in multi-billion-dollar credit portfolios.
- Managed biannual stress test of $70 billion in card loans, coordinating the process across 15 people, and producing whitepaper for the Federal Reserve.
Statistical Modeling Analyst2012 - 2012Obama for America
Technologies: Stata, SQL, Excel
- Developed finely tuned models of voter turnout and candidate support in 2012 battleground states.
- Mined massive voter-level datasets with >1,000 data fields (demographics, voting history, household info, 3rd-party data).
- Discovered prediction discrepancies and created new models which overcame poor state data in key variables (e.g. age, party registration).