Sam Magee

Performance focused software engineer.

Recent Computing graduate combining a SysAdmin background with modern Software Engineering, AI/ML, and Systems programming. Based in Austin and excited to build efficient, scalable software.

Experience

IT Service Desk Analyst @ SHI

Jan 2025 - Present

Delivering high-volume technical support for over 7,000 users with a 99% satisfaction rating. Daily operations include managing Active Directory and Entra ID access, troubleshooting issues via ServiceNow, and optimizing deployment workflows to reduce ticket resolution times by 15%.

Jr. System Administrator @ Emerald Park

Feb 2022 – Oct 2024

Administered hybrid infrastructure (VMware, Dell EMC, MSSQL) and led corporate-wide MFA rollout. Engineered custom automation tools using PowerApps and deployed a self-hosted open-source RMM solution that reduced annual costs by $15,000.

B.Sc. (Hons) in Computing @ TU Dublin

Sep 2020 - Sep 2024

Graduated with Second Class Honours, GPA 3.25. Specialized in Cyber Security & Software Engineering, bridging the gap between secure infrastructure and application development.

> Secure Programming > Penetration Testing > Algs & Data Structures > Network Forensics
Thesis: AI Content Detection

Researched and developed Machine Learning models to detect AI-generated imagery. Analyzed performance metrics against published datasets to evaluate the efficacy of current detection algorithms in a security context.

Projects

SentinelBERT — AI-powered log anomaly detection system for security events

Source
[python] [PyTorch] [HuggingFace] [Streamlit]

Transformer-based log anomaly detection system using a fine-tuned DistilBERT model to classify Windows and Azure authentication events:


> Fine-tuned DistilBERT log classifier > Synthetic + Azure VM log generation pipeline > Probabilistic anomaly scoring (LOW / MEDIUM / HIGH) > Streamlit-based analyst dashboard

InfraOracle — AI-powered LSTM based anomaly detection platform for VM's

Source
[python] [PyTorch]

SRE-style predictive infrastructure monitoring system using an LSTM autoencoder to detect anomalies in VM CPU, memory, disk, and network metrics:


> Per-VM dynamic baseline > Explainable anomaly detections > Failure risk scoring > Interactive web-based dashboard

OpsGPT — Private IT Knowledge Base

Source
[python] [LangChain] [Llama3] [FAISS]

A fully local RAG (Retrieval-Augmented Generation) chatbot for IT operations teams. Designed for:


> IT operations and helpdesk teams > SREs managing internal runbooks > Sysadmins who want searchable, conversational access to documentation

AI-Generated Content Detection using ML

Source
[PYTHON] [ML]

Research project detailing the development and analysis of various models as detection tools for AI-generated imagery.

Mechanix - Car Maintenance Tracker

Source Demo
[node.js] [supabase] [typescript] [vercel]

Car maintenance tracking platform with the following features:


> Account creation > Inter-account car transfer > Image upload for records > Virtual garage for storing multiple cars

Windows Remote Access Trojan

Source
[C] [PYTHON] [WinAPI]

Python based server side controller allowing remote control of various modules, written in C using WinAPI on an infected machine.


> Clipboard content stealer > Remote file transfer > Keylogger > Full reverse shell > Send screenshot to server > Download file to infected machine over web

Portfolio Site

Source
[HTML] [CSS] [JS]

The goal of this site was simplicity, and speed. As tested on chromium based browser, 3 requests, 5.4kb transferred, 280ms total load time. Stats on performance located in footer.

Blog

001 - The impossibility of seeing things as they are

Read
2025-12-03

On subjectivity, Umwelts, and the difficulty of human communication.