John | Security Analytics

// 01 — Work

Security Analytics Projects

LLM Abuse Detection via GPU Power Telemetry

Classifying adversarial LLM inference sessions using passive GPU hardware telemetry — no prompt content required. AUC ≥ 0.999 across 3,931 labeled sessions.

View Project → Download PDF ↓ GitHub →

Attack Triage Classifier

Two-stage classification ensemble from "attack?" to "which kind?"

View Project →

EigenFlow Profiler

Using a facial recognition approach to identifying attack types in NetFlow data.

View Project →

// 02 — Background

Experience

Cloud Platform Service Provider Oct 2017 – June 2023

Manager → Sr. Manager → Director

Recruited, hired, and mentored 3 teams of security engineers & intelligence analysts.
Created CI/CD pipeline of detections using statistical learning and traditional pattern matching techniques.
Collaborated across departments to implement secure design reviews.
Developed internal Intelligence products based off stakeholder input.
Prototyped and drove implementation of new detections during major security incidents.
Developed individual and team goals aligning with executive leadership vision.

Cloud Platform Service Provider July 2015 – Oct 2017

Lead Security Incident Response Engineer

Developed and monitored security dashboards and alerts using SIEM tools.
Conducted forensic analysis and incident response exercises.
Automated compliance checks and reporting workflows.
Led incident bridge calls for major incidents.

Several Large Enterprises July 2010 – July 2015

Security Operations Center Analyst

Monitored threats and evaluated SIEM alerts.
Automated repetitive tasks through Python scripting.
Mentored entry-level analysts on best practices.

// 03 — Stack

Skills

Technical Skills

Programming & Data Science

Python (pandas, NumPy, scikit-learn)
R
SQL (advanced querying, joins, aggregations)
Feature engineering & data preprocessing
Statistical analysis & hypothesis testing

Machine Learning & AI

Supervised learning (classification, regression)
Model evaluation (ROC, AUC, precision/recall, cross-validation)
Ensemble methods (Random Forest, XGBoost)
Dimensionality reduction (PCA, UMAP)
Anomaly detection techniques
LLM evaluation & adversarial testing (jailbreaking, red teaming)

Security Analytics & Detection Engineering

Detection engineering (rule-based & behavior-based)
Threat modeling & attack mapping (MITRE ATT&CK)
SIEM development (Splunk, Elastic / ELK Stack)
Log analysis & telemetry pipelines
Network traffic analysis
DFIR workflows and investigation techniques

Data Engineering & Pipelines

Data ingestion & transformation
Large-scale datasets (CSV, Parquet, log data)
ETL / ELT concepts
API-based data collection
Structured and semi-structured data processing

Cloud & Infrastructure

AWS (EC2, S3, IAM fundamentals)
GCP (Dataproc, Cloud Storage)
Azure fundamentals
Linux systems & scripting
Virtualization / homelab (Proxmox)

Visualization & Communication

Tableau
Power BI
Data storytelling & dashboard design
Communicating technical concepts to non-technical audiences
Interactive 3D Plots (Plotly)

Tools & Development

Git / GitHub
Jupyter Notebooks
REST APIs
Agile / Scrum workflows
ServiceNow

Focus Areas

AI Security & LLM Abuse Detection Security Analytics & Detection Engineering Anomaly Detection in Network Traffic Applied ML for Cybersecurity

Applied Experience

Built LLM red-teaming pipelines using HarmBench and adversarial datasets
Designed telemetry-driven features for detecting LLM abuse via GPU power signals
Developed classification models for cyberattack detection across NetFlow and log data
Applied clustering and dimensionality reduction to network traffic data

// 04 — Learning

Education

Online Master of Science in Analytics (OMSA)

Georgia Institute of Technology · Expected May 2026 · Program Info →

One of the most rigorous and respected online analytics programs in the country, combining statistics, machine learning, and business analytics at scale. Offered jointly by the College of Engineering and the Scheller College of Business.

Machine Learning · Regression Analysis · Data & Visual Analytics · Simulation · High-Dimensional Data Analytics · Applied Analytics Practicum

B.S. in Kinesiology

San Jose State University · 2008

Emphasis in human performance and movement science. Led to an early career interest in physical therapy through applied exercise science.

John Yater