Skip to content
View bardsleypt's full-sized avatar

Block or report bardsleypt

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
bardsleypt/README.md

Hi, I am Patrick Bardsley, applied mathematician, statistician, and machine learning dev by day, climbing bum by night.

About Me

  • I am interested in ...
    • being outside more than being inside
    • climbing and skiing
    • mathematics and physics in their full, broad, and general glory
    • quantum mechanics
    • wave phenomena
    • abstract gradient flow formulations
    • Star Wars (in order: 5, 4, 6 #EmpireDidNothingWrong)
  • I am currently learning ...
    • electrochemistry and mathematical modelling of lithium ion batteries
    • while True:
          [self.learn(val) for val in ['python', 'cool_python_packages']]
  • I am looking to collaborate on ...
    • anything with a mathematical physics aspect to it
    • anything with a statistical spin to it
    • machine learning projects (mostly signal processing)
    • anything pythonic

Contact Info

email: bardsleypt@gmail.com
resume: resume.pdf

My Portfolio

Codebase

Within this repository, I have included a fully functional codebase which can be used to train and test a neural-network-based shape-classifier. Essentially, the classifier infers the order of a polynomial with one of two architectures:

  • MLP architecture
  • Custom-kernel convolutional architecture

While this codebase and its classifier is fully functional, no efforts have been made for tuning (hyper)parameters. Rather the code is in place purely for demonstrative purposes and to collect some routines I regularly use.

A top-level jupyter-notebook is available to demonstrate the calling behavior of this codebase. By default, the notebook will execute:

  • synthetic data generation
  • a small 20-epoch training iteration
  • model calibration
  • model testing/validation

run_training.ipynb

Educational Background

  • PhD Applied Mathematics (2016), University of Utah, Department of Mathematics
    • Thesis title: Intensity-only imaging with waves, restarted inverse Born series, and analysis of coarsening in polycrystalline materials
  • Master of Science Mathematical Statistics (2016), University of Utah, Department of Mathematics
  • Bachelor of Science Mathematics (2010), University of Utah, Department of Mathematics

Professional Background

  • Cirrus Logic, Inc. (2017 - Present)
    • Senior machine learning engineer
  • University of Texas at Austin, ICES (2016-2017)
    • Postdoctoral fellow
  • University of Utah, Department of Mathematics (2010-2016)
    • Graduate research and teaching assistant

Select Publications and White Papers

Pinned Loading

  1. pybamm-team/PyBaMM pybamm-team/PyBaMM Public

    Fast and flexible physics-based battery models in Python

    Python 1.1k 558