Agent infrastructure · data systems · cloud observability

I build inspectable, testable software around AI: developer tools, agent operations, streaming data pipelines, observability platforms, and full-stack applications. I am currently pursuing a B.S. in Applied Computer Science with a concentration in Artificial Intelligence at George Mason University (expected Spring 2029).

I am open to internships and engineering opportunities in applied AI, platform engineering, cloud/security, and forward-deployed engineering.

Portfolio · Resume · LinkedIn · Email

• Applied AI engineering: I turn LLM capabilities into usable systems with instrumentation, evaluation hooks, prompt lifecycle controls, local inference, retrieval, and human-visible diagnostics.
• Systems thinking: I design clear interfaces between collectors, queues, storage, services, and operator surfaces instead of treating the model as the whole product.
• Operational ownership: My projects include automated tests, CI workflows, structured logging, health checks, deterministic behavior, security boundaries, and deployment documentation where the problem calls for them.
• End-to-end delivery: I work from backend services and data models through CLIs, terminal UIs, web applications, dashboards, and runbooks.

Go · Bubble Tea · Cobra · SQLite · GitHub Actions

• Built an agent-agnostic CLI and terminal dashboard for discovering, monitoring, and operating sessions across Hermes, Claude Code, OpenCode, Codex, and Cursor, with execution controls for Burrow.
• Designed pluggable runtime adapters, concurrent polling, deterministic session and log merging, process controls, health diagnostics, quality scoring, routing, replay, and cross-session memory features.
• Migrated the primary implementation from Python to a cross-platform Go binary while retaining the original Python test suite as a reference implementation.

Python · CLI/SDK design · YAML · SQLite · JSONL · pytest

• Agent Catalog: Declarative agent registry with schema validation, discovery, capability search, environment diffs, security metadata, and Git-friendly YAML storage.
• Agent Telemetry: Trace and span instrumentation with duration, token usage, cost, error, drift, replay, and alert queries backed by indexed SQLite metadata and full JSONL payloads.
• Prompt Manager: Versioned prompt assets with staging/production promotion, diffs, validation gates, weighted A/B tests, canary controls, and instant rollback.
• Packaged each project as a tested Python CLI with an SDK-style client and CI.

Python · Redpanda/Kafka · TimescaleDB · Neo4j · pgvector · Grafana · Docker

• Built an event-driven intelligence prototype that joins prediction-market activity with Polygon wallet behavior at the address level.
• Implemented asynchronous collectors, rolling anomaly and lead-lag signals, entity-resolution workflows, time-series storage, and a graph ontology for analyst pivots from markets to wallets, clusters, flows, and alerts.
• Containerized the services and data stack with Docker Compose and added automated linting, tests, and CI.

Proxmox · Docker Compose · Prometheus · Grafana · Loki · Uptime Kuma

• Deployed a LAN-only observability stack on an Ubuntu VM running in Proxmox.
• Unified host metrics, centralized logs, uptime checks, dashboards, and alerting; documented the architecture, installation path, security model, and operations.

TypeScript · Electron · Python · FastAPI · Tesseract · Ollama · Qdrant

• Built a local-first desktop overlay that captures on-demand screen context with OCR and routes requests to local Ollama models by default.
• Added optional cloud routing and a Qdrant-backed retrieval service while keeping those data paths disabled unless the user opts in.

JavaScript/TypeScript · Phaser 3 · Node.js · Express · Gemini API

• Created for a Gemini hackathon, combining evidence collection, suspect interrogation, and AI-assisted contradiction analysis in a playable noir RPG.
• Implemented quest and state systems, A* NPC pathfinding, responsive game UI, an API layer, and targeted gameplay/QA tests.

• NexusGate Lead Generator: Multi-step lead capture and n8n automation with enrichment, LLM scoring, personalized outreach, workflow validation, and Playwright tests.
• Portfolio Website: Responsive React, Vite, and Tailwind portfolio deployed to GitHub Pages with a custom domain.
• AI Judge Game: Python courtroom simulation using Ollama for case generation and verdicts, Kivy for a cross-platform interface, and Fernet for encrypted saved progress.
• Arch Linux Workstation Build: Documented an OptiPlex revival into an Arch Linux/Hyprland development workstation.

Browse all repositories →

• Languages: Python, Go, JavaScript/TypeScript, SQL, Bash, HTML/CSS, and C++.
• AI systems: LLM integration, agent orchestration, prompt operations, evaluation, telemetry, RAG, local models, Ollama, Gemini, and MCP/tool integrations.
• Backend and data: Async Python, REST APIs, FastAPI, Node.js/Express, SQLite, PostgreSQL/TimescaleDB, Neo4j, pgvector, Redpanda/Kafka, and Qdrant.
• Platform and DevOps: Linux, Docker/Compose, Git/GitHub Actions, CI/CD, Proxmox, Prometheus, Grafana, Loki, and shell automation.
• Web and interfaces: React, Tailwind CSS, Vite, Electron, Phaser, terminal UIs, and responsive and accessible UI.
• Cloud and security: AWS/Azure fundamentals, networking, VPN/DNS, security hardening, encrypted data handling, network isolation, and observability.

• George Mason University — B.S. Applied Computer Science, Artificial Intelligence concentration; expected Spring 2029.
• Technical Staff Intern, FBLA — Supported technical and backend operations for leadership events serving 1,200+ attendees while maintaining accurate event data.
• Intern, Science Museum of Virginia — Supported STEM exhibit and ticketing technology and collaborated on a coded planetarium project.
• Lead Lifeguard, Coastline Aquatics — Supervised teams, coordinated safety operations, and supported digital ticketing workflows.

If you are building AI products that need strong engineering around the model, I would like to hear from you. Reach me at [email protected] or connect with me on LinkedIn.