
I'm a computer scientist working on machine learning and data infrastructure. I've been building large-scale, high-performance data and machine learning systems for 15 years, with a specific focus on deep learning research and engineering for the last 8 years. I'm especially interested in how machine learning, compilers, robotics, and formal methods can make biological engineering increasingly more autonomous.
I work at Moonfire, an early-stage venture capital firm that uses software, data, and machine learning to optimize and accelerate every aspect of the venture capital lifecycle. At Moonfire, I lead technical strategy, building decision-support systems that combine data, computation, and human expertise to improve the accuracy, efficiency, and consistency of venture decisions. I also lead our internal research group, which focuses on exploring the mathematical foundations of the venture capital asset class.
While working at Facebook in 2014, I created and open-sourced osquery, which is now a foundational tool in the cybersecurity industry. Osquery is used by thousands of companies around the world to detect and respond to security threats, data breaches, and other critical incidents.
I also co-founded Kolide, served as an AI Advisor to the UK Government Office for Science, and have been a guest lecturer at ETH, LBS, and UCL.
Outside of work, I spend time in the mountains with my family and continue to develop long-standing interests in wilderness medicine and avalanche operations.
I maintain a growing ecosystem of open-source software for biological engineering across compilers, robotics, data infrastructure, machine learning, and formal methods. My focus is the infrastructure layer that can help make DBTL cycles more precise, repeatable, scalable, and increasingly autonomous, working toward a future in which biological laboratories operate more like data centers do today
Lab is a compiler for biological engineering. Scientists describe the result they want, the constraints that must hold, and the evidence needed to accept it in Python or Lab, without binding that intent to a particular laboratory, instrument, or protocol implementation. Both frontends enter the same checker, and the compiler specializes the result against the qualified capabilities of a real facility to produce reviewed work for people and instruments.
June 2020 β Present
United Kingdom
July 2022 β February 2023
United Kingdom
March 2019 β November 2020
Boulder, Colorado
May 2019 β May 2020
Montreal, Quebec, Canada
July 2016 β January 2019
Boulder, Colorado
February 2014 β June 2016
Menlo Park, California
October 2012 β February 2014
Brooklyn, New York
August 2011 β October 2012
New York, New York
January 2011 β August 2011
New York, New York
Project Lead & Founder
Release Team Member
Software Engineer
Guest Lecturer
Guest Lecturer
Guest Lecturer
American Avalanche Association (A3)
Completed Jan 2024 via Silverton Avalanche School in Ophir, CO
Canadian Avalanche Association (CAA)
Completed May 2023
National Outdoors Leadership School (NOLS)
Recertified in May 2025 in Boulder, CO
Johns Hopkins University (JHU)
Completed February 2023