AI Engineer - LLM & Agentic Systems
I build LLM-backed systems - RAG pipelines, multi-agent orchestration, and the backend infrastructure that keeps them reliable in production.
PNC Bank AI Engineer - TEKNOFEST 1st place - 3.93 GPA @ GMU
Production experience
Building fraud intelligence, agentic workflows, research pipelines, and the infrastructure behind them.
PNC Bank · Enterprise Fraud · McLean, VA
Avalanche · Seed-Stage B2B Startup · Fairfax, VA
PNC Bank · Enterprise Fraud · Pittsburgh, PA
Cuspis AI · Pre-Seed EdTech Startup · Fairfax, VA
George Mason University · Fairfax, VA
Deci Technology · Fairfax, VA
Informatik Systems · Fairfax, VA
Research & community
Accepted AI research and technical leadership across George Mason University communities.
Cinar, O. Y., Ozkose, T. E., Von Hoene, E., Roess, A., Anderson, T., and Kavak, H.
Glasgow, Scotland · Presenting author
Anderson, T., Von Hoene, S., Cinar, O. Y., Von Hoene, E., Roess, A., Crooks, A., and Kavak, H.
Glasgow, Scotland
PatriotHacks · GMU's official Major League Hacking hackathon
Oversee all company collaboration for a 1,500+ participant hackathon with 30+ big-tech sponsors, including Google, Meta, and AWS.
George Mason University Artificial Intelligence Club
Lead three end-to-end AI project teams and technical mentorship.
Selected work
AI systems built around measurable accuracy, latency, and workflow outcomes.
Some production projects are confidential and cannot be shared publicly. More public case studies will be added over time.
Featured · AI learning platform
An AI-powered lecture intelligence platform that converts raw lecture videos into editable transcripts, chapter summaries, quizzes, and targeted learning resources using Gemini-powered generation and Python video-processing services.
Fraud intelligence
Scored Zelle and ACH transactions at 0.91 recall within a 1% alert review budget, then layered a RAG-backed LLM agent over flagged transactions for analyst-ready case summaries.
Agent orchestration
Grounded agent responses over a 5,000-document corpus with FAISS/Pinecone vector search across Docker microservices, raising answer accuracy from 68% to 89% on a 120-case evaluation set.
TEKNOFEST · 1st place
A Flask and TensorFlow diagnostic application that classifies lung audio recordings for COPD risk scoring, deployed on AWS EC2 with REST APIs for hospital integration.
Technical depth
Tools I use to move AI systems from retrieval and evaluation into monitored production services.
RAG, AutoGen, n8n, Pinecone, Claude API, TensorFlow, scikit-learn, agent memory, guardrails, model evaluation, fine-tuning
Go, Scala, Flask, Node.js, Express.js, REST APIs, microservices, Teradata, Azure SQL, MongoDB, ETL/ELT, pandas, NumPy
EC2 UltraClusters, Azure, GCP, Linux, Git, model deployment, monitoring, observability, drift detection
About
I’m an AI Engineer and Computer Science B.S./M.S. student at George Mason University, concentrating in Artificial Intelligence through the accelerated BAM program.
My work spans enterprise fraud intelligence, agent tooling, model evaluation, RAG, machine-learning infrastructure, and backend systems. I also built the LLM pipeline behind two accepted 2026 Winter Simulation Conference papers.
First place, TEKNOFEST International AI Systems Competition · Google Machine Learning Bootcamp finalist from 3,000+ applicants
Contact
Connect with me about AI engineering opportunities focused on LLM systems, ML infrastructure, and production backend platforms.
