Dmitry I. Liakh (Lyakh) graduated from Karazin Kharkiv National University (Ukraine) with PhD in Physical Chemistry (Quantum and Computational Chemistry) under the supervision of Prof. V. V. Ivanov. His thesis was devoted to a systematic study of the importance of higher-order electron correlation effects for an accurate description of chemical bond breaking/formation. An indistinguishable part of his PhD work was the development and parallel implementation of the general-order coupled-cluster theory, based on a simplified variant of a domain-specific equation interpreter he developed for that purpose.
In 2009, Dmitry Liakh joined the group of Prof. R. J. Bartlett at the Quantum Theory Project (University of Florida) for a postdoctoral research focused on the development of advanced coupled-cluster methods and strategies for their massively parallel implementation. In particular, he pursued the development of an adaptive coupled-cluster theory, an approach that would ensure the accuracy of calculated molecular properties for a wide class of problems, yet with an acceptable computational cost. Besides, Dmitry Liakh published a number of papers studying the formal mathematical properties of many-body tensors. He also developed a tensor algebra library for the ACES-IV code.
In 2013, Dmitry Liakh joined the Scientific Computing group at the Oak Ridge National Laboratory. His work has been focused on enabling large-scale quantum many-body simulations on the leadership heterogeneous HPC systems. In particular, he has been working on the development of a scalable, accelerated library for hierarchical tensor algebra workloads. As part of this work, Dmitry Liakh has been pursuing the concept of domain-specific virtual processing, a potentially promising approach for complex Exascale scientific workloads. He has also contributed to a number of existing quantum-chemistry codes. Another direction of his research is directed towards enabling large-scale quantum simulations based on the tensor network theory.
R&D Activities Contributions
Center for Accelerated Application Readiness (CAAR) - In preparation for next-generation supercomputer Summit, the Oak Ridge Leadership Computing Facility (OLCF) selected 13 partnership projects into its Center for Accelerated Application Readiness (CAAR)…
Exascale Computing Project – NWChemEx - The NWChemEx project will redesign the architecture of NWChem to work with the pre-exascale and exascale computers to be deployed at U.S. Department of Energy…