Hello, I'm
Andrew Scouten
M.S. Computer Science candidate at Texas State University. I build end-to-end data pipelines and train models for real-world problems — from pavement crack detection to cancer genomics.
About
Who I Am
I'm a machine learning engineer and researcher with a focus on computer vision. My work has taken me from Texas highways (TxDOT-funded crack detection using CNNs and transformers) to low Earth orbit — segmenting corrosion in metal samples collected aboard the ISS during SpaceX-21 — with publications from both.
Outside the lab, I'm usually spending time with family, rolling dice in a D&D campaign, or tinkering with my home lab. I play a lot of video games and I'm always picking up something new — whether that's a language, a framework, or just whatever rabbit hole caught my attention that week.
Education & Experience
What I've Done
Education
M.S. Computer Science
Texas State University
Dec 2026 (expected)
San Marcos, TX
- ▸Research focus: scalable applied machine learning and domain-specific systems.
Relevant Coursework
- ▸Parallel Processing
- ▸Database Theory & Design
- ▸Image Processing & Computer Vision
- ▸Machine Learning & Pattern Recognition
- ▸Advanced Natural Language Processing
- ▸Advanced Robotics
- ▸Statistical Genetics & Bioinformatics
B.S. Computer Science
Texas State University
May 2023
San Marcos, TX
Work Experience
Graduate Research Assistant
Texas State University
May 2023 – Dec 2025
San Marcos, TX
- ▸Conducted research for the Texas Department of Transportation (Project 0-7150: AI for Pavement Condition Assessment from 2D/3D Surface Images).
- ▸Built and maintained data pipelines for large-scale pavement imaging datasets; designed, trained, and evaluated CNN and transformer-based models for crack detection, leading to a conference poster and peer-reviewed publication.
Undergraduate Researcher
Texas State University
Jan 2023 – Dec 2024
San Marcos, TX
- ▸Contributed to NASA-backed SpaceX-21 research on bacterial adhesion and corrosion using International Space Station–derived samples.
- ▸Developed data organization and quality-control pipelines; applied classical ML and deep learning methods for corrosion segmentation and labeling, resulting in a peer-reviewed publication.
Java Developer
LurgCraft, Minecraft Network
Jan 2021 – Jan 2024
Remote
- ▸Sole Java developer of LurgCraft — engineered cross-server communication systems enabling real-time synchronization across multiple game servers, Minecraft versions, and platforms.
- ▸Delivered 20 custom projects and enhanced 46 existing codebases, building strong instincts for rapidly reading and extending legacy systems.
Projects
What I've Built
- external link
OncoLearn
Hackathon · Federated LearningMulti-modal ML toolkit for cancer genomics and biomarker discovery. Led a cross-functional team of 15 contributors across three sub-projects, guiding overall architecture and development.
ClassiGraph
Hackathon · Graph MLGraph Neural Network for multi-modal cancer subtype classification. Co-developed as part of a competitive hackathon, combining graph structure with molecular feature data.
TxDOT Pavement AI
Research · Computer VisionData pipelines for crack segmentation and object detection in 2D/3D pavement surface images (Project 0-7150). Trained and evaluated CNN and transformer-based models for crack detection.
LurgCraft Network
Production · Distributed SystemsSole Java developer of a distributed Minecraft network. Engineered cross-server communication enabling real-time synchronization across multiple game servers, versions, and platforms. Delivered 20 custom projects and enhanced 46 existing codebases.
Publications
Research Output
MoPac+: Multimodal Modeling of Pavement Cracks Using Hard-Example Mining and Context-Aware Feature Aggregation
Measurement · Elsevier
Extended journal publication of the MoPac pipeline, incorporating hard-example mining and context-aware feature aggregation for multimodal pavement crack modeling from 2D/3D surface imagery collected under TxDOT Project 0-7150.
Towards Federated Learning Across Biobanks: Prototype Software from the 2026 CMU–NVIDIA Hackathon
OSF Preprints
Preprint accompanying the OncoLearn hackathon project — a multi-modal ML toolkit for cancer genomics developed across three sub-teams at the 2026 Carnegie Mellon University–NVIDIA collaborative bioinformatics hackathon.
A Blueprint for Open Science: Knowledge Graphs to Enable Biological AI Models
OSF Preprints
Documents a transatlantic collaborative effort to build and deploy knowledge graphs supporting biological AI models, developed as part of an open science initiative.
Skills
Skills & Tools
Contact
Get in Touch
Open to research collaborations, full-time ML/SWE roles, and interesting conversations (with you, yes you!). Reach out via any of the channels below.