Resume

PhD, Electrical and Computer Engineering

University of Michigan, Ann Arbor, Michigan

• GPA 3.92
• Signal and Image Processing and Machine Learning
• Co-advised by Dr. Angela Violi and Dr. Clayton Scott
• Emphasis on computational chemistry and Machine Learning theory
• Thesis Topic: Inverse Design of Nanoparticles using Small Datasets

MSc, Computer Science

University of Michigan, Ann Arbor, Michigan

• GPA 3.91
• Member of the VioliGroup computational biochemistry lab (3 semesters, 2 summers)
• President of the Machine Learning Theory Reading Group, 1 semester

BSc, Computer Science

Chapman University, Orange, California

• GPA 3.86
• Member of the Provost List, 8 semesters
• Recipient of the Chancellor’s Scholarship, 8 semesters
• Tutor and Supplemental Instructor for Computer Science and Math, 4 semesters

Hosting the Signal Processing in EECS (SPEECS) Seminar

University of Michigan

Directed Study and Summer Research

University of Michigan, Angela Violi and Clayton Scott

• Performed novel research in generalized molecular representations
• Advised computational biochemists on machine learning methodology and literature
• Supervised student researchers; Geometric Deep Learning and Deep Gaussian Processes

Graduate Research Assistant

University of Michigan, Mania Meibody

• Developed platform for remote presentation and "workshopping" of 3d architectural models using JavaScript, SASS, NodeJS, BabylonJS, and SQL
• Implemented Python scripts for automated 3D model compression
• Performed literature reviews of existing software and wrote grant proposals

Instrument Programmer

Lotus Instruments

• Developed controls for government-contracted, custom gas chromatography instruments
• Analyzed documentation to create custom libraries for serial data transfer

Software Engineering Intern

Toyoda Gosei

• Saved 2,000 man-hours and $60,000 per year through automated puchase order tracking
• Implemented a web-based asset tracking software using full-stack ASP.NET
• Collaborated with Cost Management to solidify requirements and return on investment

Linear Algebra for Data Science, Machine Learning, and Signal Processing

University of Michigan, Jeffery Fessler

• Proofread and edited textbook draft for Dr. Jeffery Fessler
• Independently verified proofs and suggested improvements for clarity and correctness
• Caught $LATEX$ typesetting errors
• Available 2024 from Cambridge University Press

The Implicit Bias of Gradient Descent on Separable Multiclass Data

• Developed a conjecture for extending existing work to include certain groups of multiclass losses
• Showed numerically that our conjecture holds for certain well-known loss functions

DJGrad: Real-Time Distributed Learning in Connected and Autonomous Vehicles (CAVs)

• Designed distributed learning protocol for sparse gradient propagation
• Implemented simulated learning environment in Tensorflow
• Demonstrated superior generalization, with fewer assumptions than Federated Learning

Domain Exploration Through Artificial Curiosity

• -Developed simulated Martian terrain as a domain - Beginning with Shmidhuber’s theoretical basis for artificial curiosity, developed an implementation using convolutional auto-encoders - Defined heuristic "Explorational Value" for evaluating path explored by model - Performed evaluation against naive models to illustrate effectiveness of artificial curiosity

Needlecast: On-the-Fly Reconfiguration of Spacecraft Flight Software

• Collaborated with NASA staff to draft specifications for protocols
• Designed a library for booting NASA core Flight System (cFS) applications on-the-fly
• Implemented Needlecast as a plug-and-play header file for NASA core cFE
• Developed a simulated network switch and web interface for straightforward debugging

AI-Driven Contemporary Archaeology for The International Space Station

• Analyzed project requirements with Dr. Walsh (co-PI of ISS Archeology)
• Compiled facial training dataset for 240 ISS astronauts
• Utilized convolutional neural networks to label astronauts’ faces in NASA photo archives