CEO / October, 2022 — February, 2024

CountOn helps people shop their values - no matter what they are. We use AI to summarize trusted product knowledge and deliver it into your regular shopping experience via a browser extension. The vision is a platform that delivers the highest quality, personalized, values-aligned recommendations into every shopping experience. In addition to being the CEO, I'm also the lead developer. I have implemented generative AI features (such as Retrieval Augmented Generation) using OpenAI, Pinecone (a vector database), and LangChain.

Senior Research Scientist / October, 2016 — January, 2022

I led sustainable customer research for five years: dozens of people in usability studies, 100's in mobile diary studies, 10k+ in surveys, 10M+ in online A/B experiments, and 100M+ in search and purchase behavior. In 2019 I helped launch The Climate Pledge and was the company's first subject matter expert on carbon offsets in support of the $100M Right Now Climate Fund. I led a project using ML to identify sustainable products and built a new sustainable shopping experience that generated over $100 million in incremental revenue. This involved collecting training data by building a Mechanical Turk task (Javascript, JQuery, CSS/HTML), preprocessing and cleaning that data (RedShift, SQL, R), training ML models (S3, SageMaker, XGBoost & other algorithms), batch processing big data, generating recommendations, and conducting A/B tests.

Behavioral Scientist / July, 2014 — October, 2016

At NREL, I introduced behavioral science research techniques to teams throughout the lab. I built an augmented reality Android app (Java, Vuforia SDK, and OpenGL ES) that incorporated carbon information into consumer good purchasing decisions and performed a randomized controlled trial at a local grocery store. The app provided personalized recommendations and led to a decrease in the average carbon footprint of purchased products and to healthier food choices. My work in smart home energy management systems contributed to a 2018 R&D 100 award (given annually by R&D Magazine, the award honors the 100 most innovative technologies of the past year). I was also involved with the design, programming, and testing of an Android app that monitors transportation related choices and suggested alternatives. This involved creating and running user-interface tests using Amazon Mechanical Turk. The experiments were coded in Javascript (with jQuery, Backbone, and Bootstrap).

Assistant Policy Analyst / September, 2010 — April, 2016

As an assistant policy analyst I navigated RAND's internal labor market. This involved networking with colleagues, getting on projects, and billing hours.