I'm an AI Engineer based in Cleveland, Ohio, with a background in Computer Science and a strong interest in building intelligent, reliable, and practical AI systems.

My work focuses on LLM engineering, agentic RAG, AI automation, and model evaluation. I enjoy building applications that go beyond simply connecting to an AI model, whether that means developing agents that interact with tools, creating systems that retrieve information from documents, or designing automated workflows that solve real problems.

I primarily work with Python, FastAPI, LangGraph, Ollama, PostgreSQL, and Docker, alongside React and TypeScript for building interactive interfaces. I also place a strong emphasis on testing, performance monitoring, and making sure the systems I develop are reliable and secure.

Outside of development, I'm continuing my education with a Master's in Human-Computer Interaction at Rochester Institute of Technology. I also hold a Bachelor's in Computer Science with a concentration in AI from Full Sail University, where I graduated as Valedictorian.

An AI document intelligence application that lets users ask questions about their PDFs and get answers grounded in the source material. It combines hybrid retrieval, local LLM inference, and LangGraph-powered retrieval workflows with page-level citations and the ability to abstain when the documents don't support an answer.

• Built authenticated document workflows with PostgreSQL/pgvector, owner-scoped access, and CSRF safeguards.
• Added Redis and Celery for background indexing, plus Prometheus/Grafana observability.
• Load-tested a read-heavy workflow with 577 requests, zero failures, and 140 ms p95 latency across 10 simulated users.

Python FastAPI LangGraph Ollama Qwen3 PostgreSQL pgvector Redis Celery React Prometheus

A local-first platform for running and comparing LLMs, recording inference history, and understanding how models and the services around them perform. I built it to get hands-on with model serving, observability, benchmarking, and reliability testing.

• Track API/inference latency and model activity using Prometheus and Grafana.
• Test database outage detection and recovery with automated integration checks.
• Benchmarked Qwen3 1.7B at 163.04 output tokens/second across 10 local Ollama requests on an RTX 4070 Laptop GPU.

Python FastAPI Ollama Qwen3 PostgreSQL Docker React Grafana Terraform GitHub Actions

An environment for testing how AI coding agents approach software engineering tasks. A tool-calling agent can inspect files, propose code changes, and run tests, while the platform records what happened and whether the proposed fix actually worked.

• Evaluates code repairs across four synthetic Python benchmarks in Docker.
• Uses fail-to-pass, pass-to-pass, and held-out tests to check fixes and regressions.
• Captures execution traces, latency, and token usage in a React/TypeScript dashboard, with 22 platform tests.

Python FastAPI Groq GPT-OSS Docker Pytest React TypeScript GitHub Actions

A human-in-the-loop support workflow that uses an LLM to classify incoming inquiries and prepare response drafts without automatically sending them. It routes sensitive or uncertain cases for review and keeps a record of what happened.

• Built authenticated webhooks, request validation, classification, and draft generation.
• Added human-review escalation, Google Sheets audit logs, Gmail notifications, and a separate error workflow.
• Validated the prototype with 3/3 API-validation checks and 6/6 classification/escalation scenarios using fictional inquiries.

n8n GPT-4o-mini JavaScript Webhooks Google Sheets Gmail PowerShell

LLMs & AI systems
Ollama Qwen3 LangGraph Agentic RAG Embeddings Hybrid Retrieval Groq API OpenAI API

Backend & automation
Python FastAPI Node.js Express n8n REST APIs Webhooks Redis Celery

Data, infrastructure & testing
PostgreSQL pgvector Docker Docker Compose Terraform Prometheus Grafana Pytest Locust Ruff GitHub Actions

Frontend
React TypeScript JavaScript Next.js HTML/CSS Figma

• Obel — A React command center for financial tracking, productivity, APIs, and AI-assisted workflows.
• Crypto Sentiment Analyzer — A Next.js/TypeScript app that combines headline sentiment and market momentum into a cryptocurrency sentiment score.
• Space Tourism Website — A responsive React interface with custom navigation and animations.