 |
Dr. Max Bloomfield
Researcher in Modeling and Simulation
|
With the pandemic, email is by far the best way to contact me reliably and quickly. I am likely to respond immediately if I am at my computer. Don't leave a voicemail on my office phone, as I rarely, if ever, check it.
In 2018 I got redirected, kicking and screaming, into machine learning and neural network research. Three months in, I had a moment of clarity in which I saw the True Power of deep learning for Bayesian approximation, and I've now jumped in with both feet. Since then I've applied deep learning-based Bayesian methods to a slew of industrially relevant problems, including microelectronics manufacturing. I am currently working on a implementation of these methods for smart manufacturing applications driven by smart manufacturing goals.
Most of my earlier publications stem from my work in microelectronics and nanofabrication models. Those include process models of reactive ion etching, heterogeneous and homogeneous nucleation of defects in SiGe devices, and finite element formulations of drift-diffusion equations for microelectronic device simulations. I'm lucky to be working for the largest (in Flops) academic computing center in the country, so I have plenty of power to work with.
For over a decade now, I have been interested in the issues that face researchers as they attempt to solve large problems on massively parallel systems. In particular, the parallelization of multiscale problems is a something that is not done well in automated ways. These challenges have driven my work in whole-device fusion plasma simulation, culminating in global impurity simulations on the ITER reactor design.
The approaches I use in my research center around Bayesian methods, variational inference, deep learning networks, level set methods, finite elements, molecular statics/dynamics, and novel iterative methods for solving integral equations.
I have been banging around the scientific community for the last decade or so (disturbingly I find a lot more 'or so' as the years go on!) doing modeling and simulation. I have spent time in industry and in academia and find the second to fit me a bit better. Although a scientist first and foremost, I find my research is almost always problem-driven, and my doctorate is in chemical and biological engineering. I do not feel that the scientific or the academic must be impractical!
I have a company through which I consult, Gnomon Logic. Gnomon Logic specializes in pre-investment technology evaluations for venture capitalists. We poke holes in other people's ideas.
I am a woodworker. A member of the Northeastern Woodworkers Association, I am becoming aware of exactly how generous and knowledgeable the woodworking community is.
Raised in NY, I am a massive Yankee baseball fan. I find baseball to be fascinating to watch and to play, though for completely different reasons.
I want to know how people think. I am interested in languages and patterns and language patterns, ideas of how people perceive and solve problems, and how we learn new concepts. I think 'innovativeness' is a skill that can be learned and practiced. As a result of this, I like to teach things in roundabout ways compared to how the are often taught.
I love the geometric. Shapes, dimensionality, origami, topology, all of that stuff.
Puns.
In an effort to actually get a page up quickly and maintain it in a useful state, I have shamelessly stolen the formating of this page from Jonathan Shewchuk's page. I find his page to be both easy on the eyes and easy to navigate.
