Mgcini Keith Phuthi

PhD Candidate

Department of Mechanical Engineering

University of Michigan

LinkedIn

mkphuthi at umich dot edu

CV

About Me

I am a Mechanical Engineering PhD Candidate at the University of Michigan at the intersection of Energy Storage, Computational Materials Science, Machine Learning and Education. My advisor is Prof. Venkat Viswanthan . My research focuses on applying Machine Learning to accelerate atomic simulations and developing methods for calculating free energies and phase diagrams. The materials I consider are materials in battery anodes such as lithium metal batteries, alloys and also metal hydrides.I was born and raised in Bulawayo, Zimbabwe and have always had an interest in science and engineering as well as methods to improve access to science education. 

Publications

Phuthi, M. K.; Yao, A. M.; Batzner, S.; Musaelian, A.; Kozinsky, B.; Cubuk, E. D.; Viswanathan, V. Accurate Surface and Finite Temperature Bulk Properties of Lithium Metal at Large Scales using Machine Learning Interaction Potentials. arXiv.org. https://arxiv.org/abs/2305.06925v1 (Accepted at ACS Omega).

Butterworth, J.; Triqueneaux, S.; Midlik, Š.; Golokolenov, I.; Gerardin, A.; Gandit, T.; Donnier-Valentin, G.; Goupy, J.; Phuthi, M. K.; Schmoranzer, D.; Collin, E.; Fefferman, A. Superconducting Aluminum Heat Switch with 3 n Ω Equivalent Resistance. Review of Scientific Instruments 2022, 93 (3), 034901. https://doi.org/10.1063/5.0079639.

Leder, A.; Anderson, A. J.; Billard, J.; Figueroa-Feliciano, E.; Formaggio, J. A.; Hasselkus, C.; Newman, E.; K. Palladino; Phuthi, M.; Winslow, L.; Zhang, L. Unfolding Neutron Spectrum with Markov Chain Monte Carlo at MIT Research Reactor with He-3 Neutral Current Detectors. J. Inst. 2018, 13 (02), P02004. https://doi.org/10.1088/1748-0221/13/02/P02004.

Research Interests

Active Research Interests:

Past research projects:

Scientific Machine Learning Webinar

I am lead organizer and host for the MICDE SciML webinar, formerly hosted as part of an ARPA-E program at Carnegie Mellon University since 2021. The SciML webinar series and panel events are organized with the goal of cross-pollinating ideas between the various emerging methods at the intersection of physics and machine learning. We have talks from speakers, mostly graduate students, on the cutting-edge of Science and Machine Learning who engage on their research in a discussion based session, chaired by leaders in the field. Feel free to see the past talks and reach out if you would like to present in the future.

Conferences and Workshops

7/10/2023: Poster Presentation: 5th iCoMSE Workshop: Machine Learning for Molecular Science. Presented on Machine Learning Potentials

3/5/2023: Presenter: American Physical Society March Meeting. Presented on the most accurate ML Interaction Potential for Lithium

8/1/2022: Attendee: IAIFI summer school on the applications of Physics in Machine Learning and Machine Learning in Physics

6/5/2022: Poster Presentation: Gordon Research Conference for Batteries on the development of a Machine Learning Interaction Potential for Lithium.

7/22/2021: Panelist: Practice & Experience in Advanced Research Computing (PEARC). Presented on work implementing Machine Learning Interaction Potentials on the NeoCortex system for the first time

8/3/2020: Attendee: SDSC Summer Institute 2020. Attended workshop on high performance computing

Press Highlights

Teaching