Jordan Donovan

Engineer Research and Development Center · Vicksburg, MS 39180 · (601) 634-4624 · jordan.t.donovanCS@gmail.com Curriculum Vitae

Responsible, committed, passionate AI / ML researcher and engineer with more than a decade of experience including deep learning and evolutionary computation research as well as desktop and mobile applications development on multiple platforms. My most recent project involves improving CNNs with perturbations to parameters at initialization which leads to avoiding vanishing gradients thus improving training efficiency and resulting accuracy of networks.


Experience

Senior Machine Learning Scientist

U.S. Army Corps of Engineers - Engineer Research and Development Center - ITL

  • Pioneered research of neural network parameters to increase efficiency of training and inference ten-fold. Projects included 6.1 Research: Intelligent Selection of Parameter Subsets.
  • Led project to design and implement an automated machine learning (AutoML) system lowering the barrier of entry for non-experts while also significantly improving the efficiency and accuracy of resulting algorithms. Projects included FLEX-4 Seed: Guided, Openended, Diverse AutoML.

2022 - Present

Senior Machine Learning Engineer

U.S. Army Corps of Engineers - Engineer Research and Development Center - ITL

  • Designed, implemented, and deployed real-time ML model for multi-modal object detection on robot increasing efficiency (by 10x) and accuracy (by 30%). Projects included Robot Engineer Operations (REO) .

2018 - 2022

Senior Software Engineer (Full Stack)

U.S. Army Corps of Engineers - Engineer Research and Development Center - ITL

2016 - 2018

Software Engineer (Full Stack)

U.S. Army Corps of Engineers - Engineer Research and Development Center - ITL

  • Implemented python, C#, C++, and JavaScript code in creating mobile applications for federal asset management increasing efficiency within this mission by 7x. Projects included Mobile Information Collection Application.

2015 - 2016

Education

University of Vermont

Doctorate of Philosophy
Computer Science

GPA: 4.0

Relevant Coursework

  • Evolutionary Computation
  • Modeling Complex Systems
  • Evolutionary Robotics
  • Deep Learning
  • Artificial Intelligence
  • Research / Dissertation - Improving Efficiency of Deep Neural Networks via Improved Initialization, Optimization, and Hyperparameter Selection

Graduate Membership

  • Neurobotics Laboratory
  • Morphology, Evolution, and Cognition Laboratory
August 2021 - Present

Mississippi State University

Master of Science
Computer Science

GPA: 3.8

Relevant Coursework

  • Graph Algorithms
  • Software Development Operation
  • Visual Data Analysis with R
  • Theory of Computation
  • Machine Learning
  • Autonomic Cyber Security
  • Reinforcement Learning
  • Research - Neuroevolution
  • Research / Thesis - Material Classification and NN Visualization
August 2015 - December 2019

University of Mississippi

Bachelor of Science
Computer Science

GPA: 3.17

Relevant Coursework

  • Computer Science I
  • Computer Science II
  • Computer Science III
  • Models of Computation
  • Systems of Programming
  • Organization of Programming Languages
  • Introduction to Operating Systems
  • Web Architecture and Programming
  • JavaScript Programming
  • Senior Project - Website Design
  • Social Responsibility in Computer Science
  • Computer Org. & Assembly Language
  • Systems of Programming
  • Software Design and Development
  • Algorithm and Data Structure Analysis
  • Multimedia Design and Development
  • Introduction to Database Systems
August 2011 - May 2015

Brandon High School

GPA: 4.0

Relevant Coursework

  • Web Design I
  • Web Design II
  • Multimedia Systems
August 2007 - May 2011

Publications

  • Donovan, J. (2019). “Understanding State-of-the-art Material Classification Through Deep Visualization.” MS Thesis
  • Donovan, J., “Perturbations to DNN Initializations Improve Optimization Efficiency and Performance by Avoiding Vanishing Gradients” (revising)
  • Donovan, J., Ramesh, D. (2024) “Inducing Diversity in Auto-ML Improves Viability” Journal of DoD Research and Engineering, FLEX-4: Semi-Autonomous Methods for Novel Neural Network Designs (submitted)
  • Hadia, X., Price, S., Donovan, J. (2023) “Semantic Segmentation: Pixelwise Classification” ERDC Library, REO
  • Donovan, J. (2022). “Brainiac+: Evolving Multiple Variables within the Brain of a Quadruped” ERDC Library, REO.
  • Donovan, J. (2022). “Innovations of Cellular Automata” ERDC Library, REO.
  • Donovan, J. (2022). “Novelty and Discovery within Cellular Automata” ERDC Library, REO.
  • Donovan, J. (2022). “Novel Feature Detectors in CNNs” ERDC Library, REO.
  • Donovan, J. (2022). “Evolutionary Selection Criteria and Performance in NAS-Bench-101” ERDC Library, REO.
  • Donovan, J. (2020). “Understanding State-of-the-art Material Classification Through Deep Visualization.” ERDC Library, REO. (Internal Only - Based on Thesis)
  • Donovan, J. (2019). “Material Classification for Robotic Integrated Engineer Operations” ERDC Library, REO. (Internal Only)
  • Donovan, J., Pettitt, J. “Mobile Information Collection Application: User Manual” ERDC Library, MICA. (Internal Only)
  • Donovan, J., Pettitt, J. “Mobile Information Collection Application: Installation Manual” ERDC Library, MICA. (Internal Only)
I realize many of these are internal only publications. These can only be made available upon specific request.

Skills

Programming Languages & Tools
  • HTML
  • CSS
  • Node.js
  • Angular
  • LessCSS
  • Python
  • Java
  • C#
  • .NET
  • GitLab
  • GitHub
  • SourceTree
  • Visual Studio
  • Visual Studio Code
  • SQLite
  • Microsoft SQL Server
  • R
  • PyCaffe
  • Tensorflow
  • Numpy
  • Scikit-learn
  • NLTK
  • Pandas
  • Matplotlib
  • Seaborn
  • PyTorch
Workflow
  • Mobile-First, Responsive Design
  • Cross Browser Testing & Debugging
  • Cross Functional Teams
  • Agile Development & Scrum
  • Data Gathering, Data Processing, & Model Building
  • Full Stack Web Development with Multiple Environments & Technologies

Interests

Aside from Computer Science, I enjoy many hobbies both outdoors and indoors. Being from Mississippi, most of the year is very warm, and I enjoy going swimming, playing soccer, rock climbing, and kayaking (the latter two of which usually involve traveling).

When forced indoors, I play video games, board games, as well as spending a great deal of my free time creating software, completing online courses, and researching state-of-the-art advances in machine learning algorithms. My aspirations include continuing my education through perusing recent research papers, completing online coursework, continuing my formal education, and experimentinig with personal projects. Most recently, I have become very interested in Deep Learning as well as nature inspired machine learninig methods. I have spent much time reading about Neurevolution of Augmenting Topologies (NEAT) and its extensions and adaptations. I have also read papers regarding swarm algorithms, state-of-the-art image segmentation, and novelty search. I have completed several exercises in data science and machine learning over the past couple of years. These can be seen on my github page linked below.


Awards & Certifications