🎯 I Build ML Systems That Find Real Edges

Production-grade AI. Sports analytics that beat the odds. Autonomous systems. Full-stack ML engineer who ships daily.

---

I'm an AI/ML engineer who doesn't just build models—I build systems that find real edges and ship to production. Whether it's a sports betting algorithm that beats Vegas, a drone that navigates autonomously, or a Web3 app that transforms NFTs into AI agents, I obsess over impact over complexity.

Currently: M.S. in AI @ USF | Founder @ Alloway LLC | OpenClaw contributor (194k+ stars) | Shipping ML daily

What I do: Turn data into competitive advantages. Build full-stack ML systems. Contribute to major open-source projects. Write about the intersection of AI and finance.

Where: Fairmont, WV (remote-first) | Status: 🟢 Open to work

Live Betting Platform | 68% Model Accuracy | Production ML

A full-stack sports betting platform with an XGBoost model trained on 5+ years of NBA data. The model finds edges against sportsbook odds using Kelly Criterion sizing. Result: Consistent value bet detection with 68% prediction accuracy. Users get daily AI picks, live odds comparison, and bankroll management tools.

Stack: React | Python | XGBoost | PostgreSQL | Stripe | AWS
Impact: 💰 Real money, real edges, real users
Live Demo | Code

Real-Time Computer Vision | 80+ Object Classes | Zero Server Processing

Built a production-grade object detection system that runs entirely in your browser. Uses TensorFlow.js COCO-SSD (MobileNet v2) to detect 80+ object classes from your webcam in real-time with animated bounding boxes, per-class confidence scores, and a live detection log. No server calls. No data leaves your device.

Stack: TensorFlow.js | React | TypeScript | WebGL
Performance: 30+ FPS on modern browsers
Live Demo | Code

Computer Vision | Path Planning | Real-Time Control

Built an autonomous vehicle platform with YOLOv8 object detection, A*/RRT* path planning, behavior trees for mission logic, and MAVLink flight controller integration. The system can detect obstacles, plan collision-free paths, and execute complex missions autonomously.

Stack: Python | PyTorch | YOLOv8 | MAVLink | ROS
Capabilities: Real-time obstacle detection, dynamic path planning, autonomous mission execution
Code

OpenClaw Integration | Multi-Asset Tracking | AI-Powered Insights

Built a financial intelligence assistant on top of OpenClaw (194k+ star AI agent framework). Tracks sports odds, NFT prices, and portfolio performance in real-time. Published 4+ custom skills to ClawHub marketplace that other builders use.

Stack: TypeScript | OpenClaw | FastAPI | React
Skills Published: sports-odds, nft-tracker, data-viz, screenshot-annotator
Code

Production API | Railway Deployed | Built by Deathconbot

A FastAPI-based AI wrapper and webhook handler that powers my automation infrastructure. Built and deployed by my own Deathconbot (autonomous AI agent). Handles AI chat endpoints (Claude/OpenAI), processes webhooks from GitHub/Telegram/n8n, and manages chat history.

Stack: Python | FastAPI | Railway | n8n
Features: AI Chat Endpoint, Webhook Handler (GitHub, Telegram, n8n), Chat History, CORS enabled
Live: deathcon-api-production.up.railway.app
Code

Power Ratings | ML Edge Detection | Portfolio Sizing

A CLI tool that finds NBA betting edges using power ratings, sigmoid win probability models, and spread + ML edge detection. Includes half-Kelly portfolio sizing and full backtesting on 2022–2025 data.

Stack: Python | XGBoost | NumPy | Pandas
Backtested: 2022–2025 seasons
Code | Live CLI

On-Chain Data | Trait-Based Personality | Smart Contracts

Transformed Mutant Ape Yacht Club NFTs into AI assistants. The system reads on-chain metadata (traits, rarity, history) and generates unique AI personalities for each NFT. Users can interact with their MAYC as an AI character.

Stack: React | Ethers.js | Solidity | OpenAI | Web3
Live: mutantintelligence.com

XGBoost Classifier | Feature Engineering | Streamlit

NBA prediction model deployed on HuggingFace Spaces. Built with careful feature engineering on player stats, team metrics, and historical matchups. Live predictions available.

Stack: Python | XGBoost | Streamlit | HuggingFace
Live Demo

Kelly Criterion | Bet Sizing | Bankroll Management

Published a zero-dependency TypeScript library for Kelly Criterion calculations. Handles bet sizing, CLV tracking, bankroll stats, and odds conversion. Used by traders and bettors.

Stack: TypeScript (zero dependencies)
npm Package | Code

Active Contributor to OpenClaw — the leading open-source AI agent framework with 194k+ stars.

• sports-odds — Real-time odds aggregation, line shopping, value detection
• kelly-criterion — Bet sizing, bankroll management, CLV tracking
• nft-tracker — On-chain NFT portfolio monitoring with price alerts
• portfolio-rebalancer — Target allocation, drift detection, tax-aware rebalancing
• market-sentiment — Fear & Greed Index, Reddit sentiment, VIX classifier
• streak-tracker — SU/ATS streak detection, regression analysis, fatigue filtering
• data-viz — Interactive charts, dashboards, real-time updates
• screenshot-annotator — Visual annotation with bounding boxes and labels
• devin-integration — Agent orchestration, task routing, failover chains

• LangChain — Robust args_schema validation for complex type annotations
• React.dev — Documentation improvements, TypeScript guide enhancements
• Kana Dojo — 15 new Japanese cultural facts with kanji

Alloway AI (Substack) — Technical writing on ML systems, AI agents, and sports analytics.

• "How I Built an ML Model That Beats Vegas (Sometimes)" — XGBoost, feature engineering, Kelly Criterion
• "Real-Time Object Detection in the Browser: TensorFlow.js Deep Dive" — WebGL, model optimization, production deployment
• "Building a Job Fit Analyzer with NLP and React" — Semantic similarity, full-stack ML
• "Sports Betting Edge Detection: From Data to Deployment" — End-to-end ML pipeline
• "Autonomous Drones Meet Computer Vision" — YOLOv8, path planning, real-time control

M.S. Artificial Intelligence — University of South Florida (2024–Present)

B.S. Data Science & Analytics — University of South Florida (2023–2025)

Roles: Data Scientist | ML Engineer | AI Engineer | Research Engineer | ML Infrastructure

Industries: Sports Analytics | Fintech | Autonomous Systems | Computer Vision | LLM Applications | Cybersecurity

What Excites Me:

• Building ML systems that solve real problems at scale
• Finding edges in competitive markets (sports, finance, crypto)
• Autonomous systems and robotics
• Open-source AI projects
• Teams that ship fast and iterate

Location: Fairmont, WV (remote-first, flexible for the right opportunity)
Status: 🟢 Available immediately

I'm always interested in discussing production ML, edge detection, autonomous systems, and building things that matter.

🚀 "Ship beats perfect. Build systems that find real edges, then share them with the world."

Last Updated: March 2026 | Status: Actively Shipping

---