Steven Jorgensen

Associate Research Staff

• Lead development of a machine learning model uncertainty quantification tool which allowed non-ML experts to understand where their ML models needed additional tuning. Tool was also a finalist in the international 2022 R&D 100 awards.
• Developed NLP malware anomaly detection algorithms to scan DNS/system logs for detection of data exfiltration over a network, resulting in a 97% detection rate on testing data.
• Developed an open-source python pypi package that allows users to easily add uncertainty quantification to new/existing pytorch models.
• Wrote monthly activity reports to keep program sponsors up to date with new research developments for 4 – 7 lab projects.
• Created and publicly released a novel dataset of 270+ hours of captured VPN/NonVPN network traffic from various computer applications for ML community use.

Research Assistant

• Developed a cutting edge software system in C++ to evolve effective, intelligent AI capable of playing strategically complex games.
• Conducted in-depth, analytical research into developing novel, efficient algorithms to evolve effective board game AI, by using a combination of genetic programming and distributed system techniques.
• Helped publish paper on useful algorithmic metrics for evaluating evolving artificial intelligence based on previous iterations.
• Implemented research tools in Python using IPython Notebooks to quickly and effectively generate meaningful graphs for analysis.

Teaching Assistant

• Assisted in facilitating an upper-level Algorithm Engineering course (Fall 2017) and an upper-level Computer Architecture course, with over 45 students enrolled in each.
• Worked with 6 - 15 students a night at office hours, assisting them in understanding topics covered in lecture such as algorithmic design, advanced data structures, and complexity analysis.
• Quickly and comprehensively evaluated student assignments and projects, and provided extensive feedback and suggestions for each individual.

Mentor

• Taught a group of 12 local high school students programming and web development through a variety of engaging activities and challenges.
• Encouraged underrepresented groups to pursue Computer Science and other STEM related majors in college through inclusive activities and guest lectures.

Massachusetts Institute of Technology

GPA: 4.0

Michigan State University

GPA: 3.79

Recipys

Created an online recipe library using RESTful text markdown and Sphinx to hold collections of favorite recipes for easy access. Also developed a webscraping script to automatically convert online recipes to RESTful text markdown. The recipe library is available at: https://recipys.readthedocs.io/en/latest/

MSU Capstone Project - "Banking with Amazon’s Alexa and Apple’s Siri"

This project extended banking options of local credit union by developing voice-controlled applications for Amazon Alexa, Apple Watch, and Android Wear. This project also implemented an administrative web portal to enable the credit union to manage the available information and services offered by these voice technologies. I was in charge of developing the Alexa application, which I implemented using AWS, Python, and POST requests.

Culinary Classification

I developed a NLP/Machine Learning software system to automate the classification of recipes by cuisine. The system was written using NLTK for the part of speech processing, Sci-kit for the ML classification algorithms, and Beautiful Soup 4 for the web scraping. I also implemented my own version for some of the ML classification algorithms, as well as a tool to calculate the mutual information provided by given features. The system was implemented entirely in Python.