SLAC National Accelerator Laboratory
Showing 71-80 of 140 Results
-
Apurva Mehta
Senior Scientist, SLAC National Accelerator Laboratory
BioI am a materials scientist with three decades of experience unraveling the molecular-scale processes that govern the functionality, aging, and failure of complex materials and devices. Over this time, advanced characterization methods have undergone a revolutionary transformation, driven by the emergence of brighter sources—from synchrotrons and X-ray free-electron lasers to MeV accelerator-based electron sources—paired with faster and larger-area detectors. While the depth and precision of measurements have vastly improved, the explosion of raw data now poses a significant challenge, making it increasingly difficult to extract meaningful insights them.
Recognizing this growing challenge, I have devoted the last decade to harnessing the power of emerging machine learning and artificial intelligence techniques to find breakthroughs. My focus has been on not only accelerating the extraction of knowledge from intricate, multi-dimensional, and often noisy measurements but also on making data collection smarter. By integrating these cutting-edge technologies, I aim to transform how we approach material science and deepen our understanding of material behavior and device performance. -
Derek Mendez
Staff Scientist, SLAC National Accelerator Laboratory
Current Role at StanfordComputational staff scientist at the Stanford Synchrotron Radiation Lightsource (SSRL), in the Macromolecular Crystallography group (Structural and Molecular Biology division).
Main activities revolve around a 3 year BRaVE-funded project to build a new X-ray resource for the acceleration of medicines development (XMeD). Active in-silico areas of research are in the characterization of crystallization outcomes, virtual and experimental screening of ligands, in-silico methods for lead molecule optimization, and extremely sensitive absorption profile characterization using machine learning techniques. A major computational goal is to determine the most valuable tools (and develop new ones) for future XMeD users.
Beyond XMeD, we are focusing on using GPUs and machine learning models to accelerate processing and characterization of user diffraction data that is collected at SSRL beamlines 12-1, 12-2, 14-1, and 9-2. In addition, we are aiming to make available to users new methods that process X-ray diffraction data at the pixel-level in order to extract more information to better resolve structural changes in proteins arising from e.g., binding events and light/chemical driven dynamics.
We are also partnering with NERSC to build a framework for our beamline users to offload computationally intensive jobs to the Perlmutter GPU cluster.
