Resume

Michael William Toomey

michael-toomey@outlook.com · michael-toomey.com · github.com/mwt5345

Download PDF

Summary

Physicist (Ph.D.) with 10+ years developing and deploying statistical and ML models for large-scale scientific data analysis — galaxy surveys, gravitational lensing, cosmological simulations. Built production-grade PyTorch pipelines (normalizing flows, diffusion models, CNNs, vision transformers), Bayesian inference frameworks, and simulation-based analysis systems. Published 25 peer-reviewed papers (1,100+ citations, h-index 15). Led cross-institutional research teams spanning MIT, Harvard, Brown, Cambridge, Edinburgh, and others; mentored 37+ students. Seeking ML engineering, data science, or quantitative research roles — preferably in Austin, TX.

Experience

Postdoctoral Research Fellow

2023 -- Present

MIT Center for Theoretical Physics

  • ▸Developed normalizing flow models to learn physically motivated prior distributions for Bayesian parameter estimation, improving inference speed by orders of magnitude over traditional MCMC.
  • ▸Built simulation-based inference pipelines combining large-scale cosmological simulations with neural network surrogates to constrain fundamental physics from DESI, BOSS, and SDSS galaxy surveys.
  • ▸Led end-to-end analysis of DESI survey data using novel statistical frameworks for model selection and Bayesian model comparison, resulting in 3 first-author publications.
  • ▸Designed and trained conditional diffusion models and vision transformers for super-resolution and feature extraction from noisy, high-dimensional image data.
  • ▸Initiated development of LLM-driven framework for automated scientific model-building; accepted to NeurIPS 2025 and the Conference on Language Modeling.
  • ▸Mentored 10+ graduate and undergraduate researchers across MIT, Harvard, and international institutions on ML-driven research projects.

Graduate Research Assistant

2018 -- 2023

Brown University — Advisor: Prof. Stephon Alexander

  • ▸Built deep learning pipelines (CNNs, unsupervised methods, domain adaptation) to detect and classify dark matter signatures in simulated gravitational lensing images; published in The Astrophysical Journal.
  • ▸Developed modified Boltzmann solvers and statistical inference frameworks to test cosmological models against observational data; key papers garnered 400+ combined citations.
  • ▸Created and released open-source software packages: NPTFit-Sim (Monte Carlo simulation), CLASS_EDE, CLASS_KINETIC (cosmological modeling), and DeepLense (ML-based image analysis).

Research Intern

Summer 2020

Microsoft Research

  • ▸Collaborated with Jaron Lanier and Lee Smolin on a research program at the interface of theoretical physics, ML, and computer science; contributed to "The Autodidactic Universe" (2021).

ML4Sci Mentor — Google Summer of Code

2019 -- Present

Google

  • ▸Mentored 25+ students across 6 years developing ML algorithms — transformers, diffusion models, physics-informed neural networks, anomaly detection — for scientific image analysis.
  • ▸Student projects resulted in 20+ publications and 16 NeurIPS ML4PS workshop acceptances.

NREIP Research Intern

Summer 2016

U.S. Naval Research Laboratory

  • ▸Developed Python-based automation tools for the Fermi Large Area Telescope (Fermi-LAT), streamlining the processing of terabytes of high-energy astrophysical data.
  • ▸Contributed to the automated identification of gamma-ray sources, directly supporting the mission's cataloging efforts.

Education

Ph.D., Physics

2019 -- 2023

Brown University

Advisor: Prof. Stephon Alexander

Sc.M., Physics

2018 -- 2019

Brown University

B.S., Astronomy & Astrophysics; B.S., Physics

2014 -- 2018

The Pennsylvania State University

Schreyer Honors Scholar — cum laude with Honors

Technical Skills

Languages: Python, C, Cython, Shell
ML/AI: PyTorch, Normalizing Flows, Diffusion Models, CNNs, Vision Transformers, Self-Supervised Learning, Domain Adaptation, Anomaly Detection, Bayesian Inference, MCMC
Data & Computing: Monte Carlo Simulation, HPC, Statistical Modeling, Image Processing, Large-Scale Data Pipelines
Tools: Git/GitHub, Linux, LaTeX, AI-augmented development (LLM coding agents); author of 5 open-source scientific software packages

Additional

  • •Peer reviewer for Physical Review Letters, Physical Review D, ApJ, JCAP, and Physics Letters B
  • •Scientific advisor for PBS NOVA's Decoding the Universe: Cosmos
  • •29 talks and seminars (21 invited) at IAS, Princeton, Cambridge, Edinburgh, Chicago, NYU, SLAC, and others
  • •2 colloquia: University of Alabama (2025), University of Winnipeg (2024)
  • •Instructor at Winnipeg Institute for Theoretical Physics Summer School (2023)
  • •Organizer, MIT Cosmology Coffee Hour Seminar (2024 -- present) & Brown Student Machine Learning Initiative (2019 -- 2023)

PBS NOVA · Decoding the Universe: Cosmos