Dr. Archis Joglekar

Dr. Archis Joglekar

ML Researcher | Research Engineer | Theoretical Physicist

Ergodic LLC

University of Michigan

Laboratory for Laser Energetics

Biography

Hi, thanks for coming. If you want to know more about me, well, this is where it goes I guess.

Short

I work on and with software at the intersection of deep learning (scientific computing more generally) and cloud computing. I am also a researcher in topics in or around so-called neural differential equations and computational physics and fusion energy. If you are still reading, and find that interesting, please reach out to me

Long

I am currently trying something new by running my own research organization that applies differentiable programming, machine learning, and numerical methods to problems in inertial fusion plasma physics. There is real promise in the synthesis of modern artificial intelligence and classical computational physics. Ergodic LLC exists to better understand that promise.

I am an affiliate research scientist with, and collaborate closely with researchers from, the Laboratory for Laser Energetics at the University of Rochester. I also am an adjunct research professor with the University of Michigan where I guide and collaborate on research on deep learning and differentiable programming for computational plasma physics.

Previously, I have worked at a couple of startups and at big tech where I have built software, performed research, hired and mentored engineers, established processes, participated in leadership meetings, and all the other pieces that go into running organizations.

I have a background in computational plasma physics. As a full-time physicist, I used supercomputers to run monolithic, massively-parallel simulation software and wrote peer-reviewed articles in journals like Nature Communications, Physical Review Letters, and others.

If you are still reading, and also thinking about things like differentiable programming, differential equations, distributed computing, computational physics, cloud computing and economics, open-source software, neural networks, scalable scientific software, please reach out to me

I also like to eat, drink, and spend time with my partner, my family, and my friends.

Interests
  • Differentiable Programming
  • Distributed, Scientific Computing
  • Applied, computational Mathematics for solving ODEs and PDEs
  • Dynamical Systems
Education
  • Ph.D. Nuclear Engineering and Radiological Sciences (Computational Plasma Physics), 2010-2016

    University of Michigan - Ann Arbor

  • B.S.E. Nuclear Engineering & Radiological Sciences, 2006-2010

    University of Michigan - Ann Arbor

Experience

 
 
 
 
 
Owner/Principal Investigator/Researcher
Ergodic LLC
2020 – Present San Francisco, California / Seattle, Washington
Needs some filling!
 
 
 
 
 
Neural Differential Equations Lead
Syntensor Inc
2022 – 2023 San Francisco, California
Needs some filling!
 
 
 
 
 
ML/AI Specialist - Partner Org
Amazon Web Services
2020 – 2022 San Francisco, California

My org works with companies that want to leverage AWS services and partners on the Cloud. I provide architectural guidance using best practices, provide technical expertise, and help connect partners to the right tools and workflows to help them in their Scientific Computing and Machine Learning journey in AWS’s ecosystem.

To continue my own learning, I develop novel applications and workflows using existing and upcoming AWS offerings in combination with Open Source Software in the scientific computing ecosystem. I also stay up to date on cloud-based machine learning and high-performance computing research and application development.

 
 
 
 
 
Adjunct Assistant Professor
University of Michigan - Ann Arbor
2021 – Present Ann Arbor, Michigan

Our team applies Deep Learning to build generative models of data collected from laser-plasma experiments. This work has been presented at the APS-Division of Plasma Physics Annual Meeting.

Our team is also developing a platform on AWS for ingesting high-repetition rate data at scale. More to come

I also serve as a guest lecturer for ENGR 151 - Computing in Engineering.

 
 
 
 
 
Head of Artificial Intelligence
Noble AI
2020 – 2020 San Francisco, California
I led the machine learning team all matters machine learning - research, development, and implementation. I also served as point of contact for ML internally and externally. I contributed as a researcher and engineer and attended conferences to distill knowledge into applications and implementations.
 
 
 
 
 
Research Scientist
Noble AI
2018 – 2020 San Francisco, California
As one of the first research scientists of an early-stage startup, I built custom deep learning layers, models, pipelines, and training infrastructure using tools like Tensorflow, MLFlow, AWS, NumPy, SciPy, Pandas, Matplotlib, etc.
 
 
 
 
 
Postdoctoral Researcher
University of California - Los Angeles
2016 – 2018 Los Angeles, California
This was very similar to my graduate school work, except I was able to use bigger and badder supercomputers like Cori @ NERSC and Bluewaters @ UIUC.
 
 
 
 
 
Graduate Researcher
Los Alamos National Laboratory
2014 – 2014 Los Alamos, New Mexico
Investigated implosion dynamics of inertial fusion experiments to discover the effects of viscosity. Published peer-reviewed journal articles.
 
 
 
 
 
Graduate Researcher
University of Michigan - Ann Arbor
2010 – 2016 Ann Arbor, Michigan
I performed research on computational plasma physics with high-performance, accelerated computing. This involved using supercomputers (1000+ CPUs and GPUs) to perform massively-parallel simulations and large-scale data processing. I was also invited to participate and present at conferences and workshops, and published peer-reviewed articles

Recent Publications

(2023). Machine learning of hidden variables in multiscale fluid simulation. Machine Learning: Science and Technology.

Cite DOI URL

(2022). Unsupervised discovery of nonlinear plasma physics using differentiable kinetic simulations. Journal of Plasma Physics.

Cite DOI URL

(2021). Observations of pressure anisotropy effects within semi-collisional magnetized plasma bubbles.

Cite DOI URL

(2020). Observations of Pressure Anisotropy Effects within Semi-Collisional Magnetized-Plasma Bubbles.

Cite URL

Contact