Levi is a graduate student researcher in the Department of Electrical and Computer Engineering at Texas A&M University in College Station, TX. Additionally, he is a member of the United States Army Reserves, currently serving in an Aviation Battalion in Conroe, Texas as a Blackhawk pilot. He is a Research Assistant in the Department of Electrical Engineering at Texas A&M and is currently collaborating with the Army Research Lab on research areas of interest to the Department of Defense.
He holds a Bachelor and Master of Science in Electrical Engineering from Texas A&M, completed in 2015 and 2016 respectively. He was formerly a member the Corps of Cadets, and while pursuing his Masters served as the Executive Vice President of Texas A&M’s Student Government, sitting on numerous university advisory boards and committees and helping found the Texas A&M University Career Closet.
In 2019 he was appointed to by the Governor of Texas to sit as Student Regent to the Texas A&M Board of Regents and later as the student representative to the Texas Higher Education Coordinating Board.
He has attained numerous civilian and military awards, including the Buck Weirus Spirit Award for his service to the Texas A&M campus community and the distinction of Distinguished Honor Graduate of his Air Assault class.
Levi’s Research Interests are in artificial intelligence with applications in material systems, and lie in computer vision applications and at the intersection of deep learning and physics. He has held many fellowships as an undergraduate and graduate student:
PhD in Electrical Engineering, 2021
Texas A&M University
M.S in Electrical Engineering, 2016
Texas A&M University
B.S. in Electrical Engineering, 2015
Texas A&M University
Skills involving a keyboard
Extensive scripting experience in Python
Experience building and training deep learning algorithms Tensorflow and Keras
Organizer of the Julia subgroup in the Packt Reviewer Committee
Wrote and published the open-source software package R:BoolFilter to CRAN
Development skills specifically for multi-GPU training
Development and model training experience with Nvidia GPU Support in deep learning AMIs
Skills not related to a keyboard
Experience running organizations of 5-50+ people
Member of the US Army Reserve Component for ~10 years
Hundreds of flight hours in the UH-60 and other DoD and civilian airframes
Sat on numerous advisory and executive-level boards
Years of experience of advocacy in the higher education sector
Professional experience in the research and academic realms
Research engineer in the Vehicle Technology Directorate (VTD) applying multi-GPU computer vision and deep-learning based generative models to microscale material images, bridging the gap between GPU-enabled deep learning and what we understand about material systems in the physical world
Additional projects include: bringing domain expertise into a large-scale project for the Joint Artificial Intelligence Center (JAIC) involving analysis of over 27 TB of health monitoring system data for the H-60 platform - utilizing Nvidia’s RAPIDS platform for bayesian methodologies as well as recursive deep learning
Ph.D. Student at Texas A&M University studying machine learning, deep CNNs, adversarial networks, and materials informatics.
Currently has many publications in multiple journals and IEEE Conference Proceedings, and has also published and currently maintains the open source R software ‘BoolFilter’ on the worldwide R repository, CRAN.
Note: In the US Army, pilots who are also commissioned officers hold leadership positions within their units, in addition to being a rated pilot and performing flight duties.
Awards: Distinguished Honor Graduate, Air Assault Course, #1 of 220
Directly responsible for almost 30 personnel and 5 aircraft, totaling approx. $110 million in Army assets.
Job requirements consist of maintaining flight minimums, training requirements, and professional military education while enabling other personnel in the platoon to do the same.
Flight school student in the IERW UH-60M track at Ft. Rucker, AL.
Graduated as the top ranked reserve component officer and second overall in the class of over 25 commissioned and warrant officers.
Volunteer experience and notable student involvement
The mission of the Texas Higher Education Coordinating Board (THECB) is to provide leadership and coordination for Texas higher education and to promote access, affordability, quality, success, and cost efficiency through 60x30TX, resulting in a globally competitive workforce that positions Texas as an international leader.
Appointed by the Governor of Texas to advocate on behalf of the ~1.5m students in institutions of higher education in the state of Texas. Responsible for advocating for affordability, experience, and quality of higher education institutions and organizations.
Appointed by the Governor of Texas to serve as the sole student member to the Texas A&M Board of Regents - responsible for advocating for ~167,000 students across 11 campuses
Directly provide oversight to the Texas A&M University System’s budget of ~$5 billion in revenue and expenditures, reviewing proposals as fit for the individual campuses in the system
Responsible for stepping in for the Student Body President at a moment’s notice and coordinating the efforts of the Student Government Executive Cabinet to align with the SBPs vision.
Sat on numerous university committees and boards, including Recreational Sports, Student Affairs Fee Advisory Board, Student Organization Funding Assistance Board, Campus Concealed Carry Task force, and others.
In many real-world applications of deep learning, estimation of a target may rely on various types of input data modes, such as audio-video, image-text, etc. This task can be further complicated by a lack of sufficient data. Here we propose a Deep Multimodal Transfer-Learned Regressor (DMTL-R) for multimodal learning of image and feature data in a deep regression architecture effective at predicting target parameters in data-poor domains. Our model is capable of fine-tuning a given set of pre-trained CNN weights on a small amount of training image data, while simultaneously conditioning on feature information from a complimentary data mode during network training, yielding more accurate single-target or multi-target regression than can be achieved using the images or the features alone. We present results using phase-field simulation microstructure images with an accompanying set of physical features, using pre-trained weights from various well-known CNN architectures, which demonstrate the efficacy of the proposed multimodal approach.