About Tim Hourihan
I'm a Research Programmer at Boston Fusion, where I integrate and adapt
machine learning models for computer vision applications in national defense.
I work with architectures like transformers and joint embedding predictive
architectures (JEPAs), modifying them to meet our specific needs.
Before joining Boston Fusion, I graduated from the University of Rhode Island
in 2024 with a Bachelor of Science in Computer Science and a minor in Data
Science, earning a 3.99 GPA. While in school, I worked as a Software Developer
at OMNI Digital Services, building full-stack websites and internal tools with
JavaScript, PHP, and MySQL. I also completed a Software Development Co-op at
FM Global, contributing to a large-scale internal web application using React
and C#.
Outside of work, I’ve contributed to a range of academic and personal projects
involving machine learning, software development, and game design. Some of my
favorites include a deepfake detection system using EfficientNet and FasterViT
that placed first in a machine learning course competition and a sentiment
analysis visualization built during HackMIT, where my team earned third place
in the MIT-IBM Watson AI Track. I’ve also published several games on itch.io
using Unity and Godot.
I truly enjoy solving challenging technical problems and I'm always interested
in learning more and growing as a developer in this field.
Tim Hourihan's Work
Jan 2025 - Present
Boston Fusion - Associate Research Programmer
- Implementing and optimizing machine learning algorithms with applications to national defense.
June 2024 - Jan 2025
OMNI Digital Services - Software Engineer
- Introduced GitHub version control and a CI/CD pipeline for automatic deployment to a remote
server with FTPS, significantly improving code management and deployment efficiency for the
company.
- Utilized Python scripts to automate several redundant operations relevant to the job
- Created a zip-code-qualifier tool that leverages OpenAI's GPT API to discover cable/satellite
providers offering services for specific zip codes the user enters
January 2024 - June 2024
FM Global - Software Development Co-op
- Worked on a large full-stack internal application, the Digital Client Experience Location Risk
InSite platform, performing in an agile scrum environment with other developers using React,
TypeScript, GraphQL, Hasura, C#, and SQL.
- Created scalable APIs for a GraphQL layer of the application leveraging .NET, DTOs, EFCore,
FluentValidation and AutoMapper.
May 2022 - January 2024
Omni Digital Services - Junior Software Engineer
- Developed a “Project Starter” web application that streamlined the creation of frontend,
backend, and database architecture for any web project.
- Constructed several full-stack websites for local businesses with jQuery, PHP, MySQL.
- Created a custom web-based inventory management system for a traffic distribution company,
automating the building of complex PDFs for BOMs and POs for 400+ vendors, 700+ customers, and
1700+ parts. Used PHP, FPDF, and RESTful APIs.
June 2020 - September 2020
PrepareRI - Research and Development Intern
- Collaborated with a team of 25 to prototype a wearable device and mobile application for PURVIS
Systems that allowed US military personnel to identify pre-symptomatic COVID-19 cases, and
achieved first place in the event out of 20 teams.
Tim Hourihan's Projects
Deepfake Detection App
September 2023 - December 2023
Trained EfficientNet and FasterViT models to classify deepfake images, achieving an accuracy of
90%. Integrated EfficientNet with Grad-CAM to display the most active tensors of the image and
created a Gradio interface. Scored first place in the project out of 14 teams of undergraduate
and graduate machine learning students.
HackMIT “Medisum” App
September 2023
Developed a scalable web app that displays 50+ hospitals in the Greater Boston region with scores
based on the sentiment analysis of their reviews and provides a summary of those reviews on an
interactive map. Built a backend service using FastAPI and the Yelp Fusion API. Leveraged
Hugging Face's BART Large CNN model to generate summarized reviews and IBM Watson's sentiment
analysis to analyze emotions in text. Finished in third place at HackMIT in the MIT-IBM Watson
AI Lab Challenge focused on combining AI and the Physical World.
University of Rhode Island Web App
October 2023 - December 2023
Prototyped a character creation tool that analyzes young students’ perceptions of scientists by
collaborating with the Computer Science and Education departments at URI. The project was funded
by the NSF (National Science Foundation).
Rust Image Compression Tool
March 2024 - April 2024
Worked with a partner to implement a lightweight version of JPEG in Rust based on specifications
by programming a CLI tool for Portable Pixmap (PPM) image compression and decompression.
Employed floating-point quantization, discrete cosine transformation, and bit packing for a
binary reduction size of 67%.
Game Development Projects
Various Dates
Developed multiple games for game jams on itch.io using Unity (C#) and Godot (GDScript).
Contact Tim Hourihan
Feel free to get in touch with me, I look forward to meeting you soon.