Sam Zimmerman Email and Phone Number

PathCheck was an unprecedented spinout from MIT that built cutting-edge exposure notification and digital pandemic response open source software to help end the COVID-19 pandemic. PathCheck was organized and operated like a software company. As CTO, I led a remarkable team of more than 400 elite professional volunteers from the software industry, healthcare, and academia to create mobile solutions deployed by 9 states and countries to more than 5 million users.

Sagewell was a seed-stage startup I founded aimed at serving the financial needs of 55+ Americans. We served thousands of members and raised 6.5M.

Consulting with Rama Nemani and Thomas Vandal on applications of computer vision to climate science.

Freebird empowers travelers to skip the line and instantly book a new ticket – on any airline at no additional cost – in the event of a flight cancellation, significant delay, or missed connection.In this role I recruited and managed a full-stack 15-person tech team (engineering, product, design, security), supported fundraising ($16.5M), and led our data science and risk team. The average tenure of the tech team was nearly three years.

At the Tenenbaum Lab, I am named on a grant through the Templeton Foundation modeling how people make decisions about things they've never experienced before. To this end, I model these decisions using probabilistic programming methods (in CHURCH, STAN, and WebPPL) to instantiate agents that reason using structured object representations represented via hierarchical nonparametric Bayesian clustering techniques to model the utility of a new experience. I also conduct experiments with people to validate these models.I am also exploring the extent to which people reason using second-order decision theory. Would a Red Sox fan rather be a Yankees fan, or not like baseball at all? How do people approach these decisions? Can we model those decisions algorithmically?

At PowerAdvoate, I was a technical lead in a multi-team effort to automate large scale data analysis techniques currently done at our company -- like mint.com for energy companies. Here I was responsible for technical coordination between the development team, the data science, and the data analyst teams in this large scale automation project.This involved data analysis and algorithm prototyping in Python and then assisting in the implementation of these models in Scala and Java using the Berkeley Stack (SPARK and MLLib). We then served this model in a J2EE webapp I helped develop, which also stored and structured several proprietary ontologies in the energy sector our teams created.

At Boston Technologies, I was part of a three-person team responsible for hedging our e-commerce FX risk, and at times our commercial risk using our closed-form pricing engine. This involved deploying various quantitative analysis techniques to optimize parameters for trading and hedging strategies in Python, Java, and R. Before deploying those algorithms, we had to create a large scale backtesting frame using Java and Python. Additionally, our team built graphical user interfaces to aid in monitoring this flow.

At Evoapp, I was exposed to and built (with a coworker) my first machine learning models to analyze the sentiment of tweets, blog posts, facebook posts, and scraped websites. This model was later served in production and anchored our analytics platform. I also developed weighting algorithms to better account for natural language changes in Tweets.

In the summer of 2009 I worked to manage Duke Wrestling's recruiting plan as well implementing various technology updates to recruiting tools, video streaming, video editing, and budgeting.