Turning the house I grew up in into a fishing retreat on Lemon Bay.

Providing technical oversight, system architecture design, and custom development services for high-throughput computational systems, particularly computer vision applications.Select clients: Marble Technologies, Cognito Health, Aerofarms, Natura, Essess, Gestsure, Mind's Eye (DARPA), Mindboggle (NIH)

I spun FØCAL out of my consulting practice in 2015 to focus on problems that I was seeing again and again when inexperienced teams tried to build computer vision systems. I bootstrapped for a year before turning to venture capital. I raised ~$3M in pre-seed capital as a first-time, lone founder.FØCAL built tools for bringing hardware acceleration to software people, especially in edge compute applications: (1) a device farm IaaS -- "AWS for hardware accelerators" -- and (2) instrumentation for understand your architecture's performance, on or off the farm.

I was the first hire at LM-ATL with a degree in a computational science and, through outstanding performance on numerous DARPA research contracts, helped it to grow the multi-million dollar Distributed Processing Lab. Throughout, I applied my capability as scientist and a talent for systems engineering to building production-ready prototypes that solved mission-critical problems in a variety of machine learning domains. The performance constraints supplied by our customer were unsurpassed, forcing us to effectively integrate bleeding-edge academic research and develop our distributed learning systems on HPC platforms, often directly in hardware. Both with the customer and amongst my peers I built a reputation for exceptional problem specification, compelling system analysis through simulation and visualization, and an overwhelmingly effective prototype-to-production development process.

With Drs. Ennio Mingola & Gail CarpenterDepartment of Cognitive & Neural SystemsI worked with senior faculty to developed new models of primate visual cortex for applications to image processing in remote sensing tasks. Our efforts were funded by a grant from the National Geospatial Agency (NGA, then NIMA) via DARPA. The goal was to build a computer vision system capable of identifying potential groundwater sources from SAR imagery.With Dr. Tommaso ToffoliDepartment of Electrical & Computer EngineeringThe goal of Boston University's EarLab project (http://earlab.bu.edu) is to build a distributed auditory modeling environment, a simple interface to an HPC system capable of executing high fidelity and massively parallel simulations of neural auditory processing. I worked with senior faculty to provide a concise model specification language -- a "dialect" of Python --­ for constructing large-scale models, collecting data, and streamlining analysis.