resume

Research Software Engineer

MathWorks

Development in Deep Learning, Image Processing, Computer Vision, and Automated Visual Inspection areas

• Performed literature surveys and implemented state-of-the-art research into computer vision products.
• Fine-tuned a multimodal embedding model for use in industrial anomaly classification and segmentation tasks.
• Investigated compression methods for physics-informed neural networks on edge devices.
• Shipped 5 image metrology tools with proprietary snap-to-edge functionality for automated measurements.
• Implemented industry standard template-matching algorithms from expired Cognex PATMAX patent.
• Validated performance of a new loss function for a 3D pose estimation model and designed an interface for it.
• Trained and coached colleagues and customers on deep learning in MATLAB, PyTorch, and TensorFlow.

Robotics Software Engineering Intern

Amazon Robotics, Innovation Lab

Designed, architected, and implemented C++ software for a semi-autonomous, mobile manipulation robot.

• Wrote synchronized perception (OpenCV), motion planning, and control orchestration microservices in C++.
• Collaborated with mechanical and electrical engineers as part of a small research group.
• Ran live robot demos for Amazon Robotics Leadership and external stakeholders.

Data Science Intern

Tesla, Energy Reliability Team

Led the Energy Reliability team’s transition to a proprietary hardware test automation framework in Python.

• Debugged multi-threaded Python programs to optimize framework’s performance for long-running tests.
• Completed ad-hoc data analyses on Tesla’s fleet telemetry using Python and Spark for various stakeholders.

Security Engineering Intern

Dell Technologies, Hardware Security Team

Developed a cryptographically secure supply-chain component verification solution for Dell storage platforms.

• Built a RESTful service using Python, Flask, SwaggerUI, and MongoDB as a proof of concept.

Master of Science

University of Massachusetts Amherst

Computer Systems Engineering

Thesis:

A Model Extraction Attack on Deep Neural Networks Running on GPUs

Bachelor of Science

University of Massachusetts Amherst – Commonwealth Honors College

Computer Systems Engineering, Minor in Philosophy

Thesis:

Toward Mitigating Adversarial Attack on Neural Networks via Hybrid Neural Network Decision Tree Architecture

2nd Place - New England Hardware Security Conference Poster Competition

New England Hardware Security Day

Gave a research presentation on Master's Thesis: 'A Model Extraction Attack on Deep Neural Networks Running on GPUs'

Departmental Three Minute Thesis winner

University of Massachusetts Amherst

Departmental Three Minute Thesis winner for presenting my research to a nontechnical audience.

Extracting DNN Architectures via Runtime Profiling on Mobile GPUs

IEEE Journal on Emerging and Selected Topics in Circuits and Systems

Kim, Dong Hyub, Jonah O'Brien Weiss, and Sandip Kundu.

EZClone: Improving DNN Model Extraction Attack via Shape Distillation from GPU Execution Profiles

arXiv preprint arXiv:2304.03388

Weiss, Jonah O'Brien, Tiago Alves, and Sandip Kundu.

Hardening DNNs against Transfer Attacks during Network Compression using Greedy Adversarial Pruning

IEEE 4th International Conference on Artificial Intelligence Circuits and Systems

Weiss, Jonah O'Brien, Tiago Alves, and Sandip Kundu

Baseball Umpire Assistant

Engineered a networked embedded system to aid baseball umpires in making calls at first base with 25ms accuracy via high resolution timestamps and clock synchronization between devices. Wrote in C and Python.