Many companies offer data science capabilities but focus their efforts on areas like web traffic analysis, advertising, or general statistics. Protection Engineering Consultants (PEC) offers the unique ability to work at the intersection of advanced engineering problems and cutting-edge data science methods.
Our mission is simple: to apply advanced data science tools to first-principles engineering and physics problems, an intersectional domain we refer to as Engineering Analytics.
Our team has long experience in advanced engineering research, which has always leveraged tools such as multi-dimensional optimization, stochastic and probabilistic modeling, and design of experiments. During the past several years, our work has continued to expand into new areas as demanded by novel and challenging problems. Today we actively apply techniques such as computer vision, time series and frequency domain analysis, symbolic regression, evolutionary algorithms, uncertainty quantification, and deep learning. Yet our essential paradigm remains unchanged: to use advanced data methods to accomplish the core mission of solving first-principles engineering and physics problems.
We help our clients navigate the challenging intersection of engineering, physics, and data science. Fields such as machine learning, uncertainty quantification, and computer vision become more complex each year. Upline management in government and industry are increasingly requiring that solutions leverage the most advanced techniques possible. This demand presents challenges to program managers, as it can be difficult to understand which algorithms would add value, and which would simply add noise. We help develop strategies for our clients to obtain robust first-principles solutions.
Real-time calculation of threat risks requires fast-running models, which are often developed using large amounts of test data or finite element model data. As threat scenarios become more complex, such models become more complex and operate over higher dimensional spaces. Industry analysts have increasingly turned to black box machine learning techniques, but the resulting models are often difficult to interpret and can lead to fragile performance. At PEC, we develop fast-running models following a physics-based paradigm, which allows our models to be transparent and interpretable.
Advanced Testing Analytics
PEC employs machine learning, computer vision, and signals analysis methods to advanced testing and measurement. Challenging engineering and physics tests often hinge upon the ability to measure key phenomena. Our intersectional strengths in engineering and data science enable us to develop novel solutions and successfully make key measurements. We develop algorithms to measure things such as detonation events, shock physics impacts, and material properties. Our goal is to squeeze every drop of information from the measurement data, while quantifying the data limits and level of uncertainty.