Kieran McNulty
Hi, I'm
I am a Fall 2024 Computer Science grad from Cal Poly, San Luis Obispo.
I have a specialized focus and drive for Embedded Software and Linux Development. I also have interest areas in security and privacy, machine learning, and data science
General Atomics ASI
Embedded Software Testing Summer 2022
In my time at GA I created an automated testing tool to simulate flight scenarios for an embedded system on General Atomic's unmanned air vehicles.
This testing process involed turning flight scenarios from test cases into input from a ground control station and sensor the ground control station to the flight computer, and recording and analyzing the flight computer's behavior against test cases.
Working with a small team, we took this project through the software development life cycle from requirements gathering, design, and implementation, complying with rigorous certification guidelines.
Internships
Wind River Systems
Linux Build Systems Development Summer 2023
My internship at wind river was a deep dive into linux build systems, working with the Yocto Project and Open Embedded build system.
I resolved critical bugs in a tool that automates the Linux image building process and makes it user friendly to create a custom image for a user's machine and software requirements.
I also got experience interfacing with the Yocto Project open source community, getting multiple improvements to the project pushed upstream.
Membership Inference Attacks on Federated Learning
Following a research paper on a novel method of membership inference and source inference based on bias gradients, I created my own implementation and tests on my own Federated Learning Simulation.
I conducted attacks posing as a malicious participant, malicious aggregator, and a man-in-the-middle, achieving accuracy in all scenarios, and conducted experiments to learn what FL implementations are vulnerable to membership and source inference attacks.
File Search
Implemented file searching for a linux device in C++, using pipelining and pipe redirection, signal handling, child process management, and interrupt service mechanisms
High Performance Programs
Designed an efficient controlled parallel program for matrix multiplication and processing image files in C++ utilizing MPI with shared memory for distributed processing
Recent Projects
C Compile-Time Stack Smashing Protection
Created an LLVM pass to analyze intermediate-representation code for buffer overflow and stack smashing vulnerabilities, using both static and inter-procedural analysis.
Around these vulnerable areas unique stack canaries are placed, so that after writes we can check to see if the canary was overwritten and abort the program
Just for fun Website Links
PCPricer.net is a a data science tool for estimating the value of used PC components, as well as creating and sharing PC builds.
A tool to see a breakdown of your spotify listening based on gender, for awareness or just curiosity
