This page presents other evidence of my professional capacity in computer science as required for the ICS Ph.D. Portfolio. This includes all forms of research, teaching, and service to the community and discipline, including professional vita of employment, professional presentations, reviewing of papers for conferences and journals, competitive fellowships or other external funding awards, patents, teaching, service on committees or as graduate student representatives, and letters of reference.
In my current role I lead a team of approximately 25 full-time staff in the development and fielding of machine intelligence and autonomy solutions to space domain awareness challenges. Our work includes global, autonomous sensor management capability development, applied computer vision for scientific imagery exploitation, and digital infrastructure software development.
The Machine Intelligence for Space Superiority (MISS) lab was founded in 2017 to advance the state of the art in computer vision tasks related to space domain awareness. Members of this lab have published dozens of papers on topics ranging from non-resolved spectroscopic object recognition to extended object imagery interpretation. MISS has received approximately $15M of external funding since the group was founded in 2017. Funds sources include core and external funding, as well as two Air Force Office of Scientific Research (AFOSR) grants ($400K total). For my work leading the DOTS program and creating the MISS lab I was awarded the 2019 AFRL Commanders Cup, the most competitive annual award available to any of AFRL’s 5,400 employees.
As the Chief Engineer of the Maui High Performance Computing Center (MHPCC) I authored several strategic plans, acquired high-performance computing (HPC) systems, and supervised technical personnel. During my time at MHPCC, I served as the DoD High Performance Modernization Program liaison to the USINDOPACOM joint staff for HPC and machine learning applications.
The Dynamic Optical Telescope System (DOTS) was an Air Force Research Laboratory advanced technology demonstration (ATD). As the program manager, I was responsible for the execution and success of this $33M ATD, which included large-scale, multi-system autonomous software development, novel charge-coupled device manufacturing, and optical system design and integration.
Supervised serval scholars during 3-month summer internships from 2017-2021. Past scholars include:
My role as the lead for the USSF SSC/ECZGA AI and Autonomy project, I directly supervise a variety of research and development projects. Three of these projects are described below.
SatNet is a learned approach to deep space resident space object detection. This approach has been replicated several times in many defense and academic research groups for a variety of ground- and space-based telescopes. SatNet has spawned 17 annotated datasets, each from a different sensor to enable domain adaptation, as well as a variety of subsequent publications that either improve on the original performance or present alternative approaches using the same methodology. The SatNet approach to deep space object detection is currently planned for adoption in two separate Space Systems Command programs of record and will be used in the upcoming Sprint Advanced Concept Training space warfighting exercise to perform real-time satellite breakup analysis.
SpectraNet is a learned approach to deep space object recognition using spectrograph imagery. Satellite re-identification is essential for accurate space domain awareness but can be very challenging when satellites come close to one another. Spectral exploitation for satellite recognition has been attempted for two decades, but has never succeeded on-sky (i.e., no classifier has ever generalized to real observations). We hypothesized that this was due to noise introduced during the calibration and reduction process, and instead sought to learn a direct mapping from spectrograms to object identity. We successfully demonstrated on-sky recognition at the Maui Space Surveillance Complex in late 2021 and are currently maturing and transitioning this capability to routine operations.
DASIE is a technology maturation program for distributed aperture optical interferometry. This program involves the joint design of a distributed aperture telescope system and subaperture articulation policies to realize cost-effective large-aperture imaging. An optical bench demonstration of the telescope system is nearly complete at the Maui Space Surveillance Complex, and the first open-loop articulation policy design paper was published in early April.
My work requires routine professional presentations. Limiting to flag officers or large audiences (>50 attendees) these are too numerous to list (~25/year), so I’ve selected a few that stand out as exceptional. This list excludes all archival conference presentations, which are identified in my evidence of scholarly ability.
Artificial Intelligence for Space Situational Awareness. JASON Invited Talk, July 11th, 2017. San Diego, CA.
Autonomous Dynamic Sensor Management. JASON Invited Talk, July 27th, 2017. San Diego, CA.
Prospects and Progress in Artificial Intelligence and Autonomy for Space Warfighting Applications. United States Air Force Scientific Advisory Board, May 21st, 2021. Virtual.
Digital Infrastructure and Machine Intelligence for Satellite Collision Analysis. United States Air Force CORONA TOP, October 3rd, 2020. Virtual.
Node-Level Deep Learning Input Pipeline Optimization on GPGPU-Accelerated HPC Systems GPU Technology Conference, March 28th, 2018. San Jose, CA.
Dynamic Optical Telescope System Advanced Technology Demonstration. United States Air Force Scientific Advisory Board, November 16th, 2017. Albuquerque, NM. Best Presentation
Session Chair, Space Domain Awareness, SPIE Defense and Commercial Sensing, SPIE. April 2022.
Session Chair, Artificial Intelligence, Spaceflight Mechanics Meeting, American Astronautical Society. February 2021.
Area Chair, Machine Learning for Space Situational Awareness, AMOS Conference. September 2020.
More details can be found on my CV.