Brian Maddock Email & Phone Number

AI/ML Research Scientist with four years experience in applying machine learning to ISR problems. Background in theoretical model building, machine learning, computer vision, and overhead sensing including EO and SAR modalities, as well as over twelve years of experience in high level research in theoretical, experimental and computational physics. Passionate about solving challenging data-informed problems, and continuous improvement as a researcher, musician, and athlete.Work Experience: - Improved the performance of classification algorithms via the introduction of neural aggregation networks to apply attention to temporal data sequences in ISAR imagery, and by utilizing momentum contrastive learning to exploit large unlabeled ISAR imagery dataset to create high performing pretrained models.- Principal Investigator on a task order which designed and trained detection algorithms for use on SAR imagery. This included training and evaluating detection algorithms, secondary classifiers, complex valued networks, implementing SAR pre-processing techniques, and custom SAR augmentations, and deploying systems in operational environments. - Provided ML expertise to train and evaluate model performance to understand the applicability of using synthetically generated SAR chips to improve model performance on obscured targets. Synthetic chips were developed using a variety of approaches including xPatch and generative adversarial networks. - Developed experiments and performed studies to determine the ability of machine learning algorithms to perform specific NIIRS level tasks during overhead detection, and chip level recognition in SAR imagery. This was done as a function of type information/square meter to determine the quality of imagery. - Principal investigator on Phase I effort which implemented autoencoders based approaches to address defect detection during additive manufacturing processes using laser powder bed fusion machines.- Assisted mechanical engineers in developing an ML based controller for a 3DOF system.- Primary author and principal investigator of several successful proposals (Phase I, Phase II, and IDIQ Task Orders).- Organizer of in-house weekly forum, and publisher of weekly ML blurb focused on recent developments in ML.

• 

  Research Scientist

  Etegent Technologies

  Mar 2020 - Present