Trading and investing systems + Cloud engineering. I build AI agents, financial data tools, and the infrastructure they run on, with a focus on turning messy real-world data into decisions.
- Focus: quantitative research, LLM agent systems, and automation
- Daily driver: Python, with modern AI tooling
- Deploy with: Docker and cloud-based workflows (AWS Cloud Practitioner in progress)
- Security-minded: on-device processing, OAuth, and careful handling of tokens and secrets
- Studying: Computer Science @ CSU San Marcos
- retro-spotify-voice-remote — An offline voice remote for Spotify Premium. Speech is processed entirely on-device with Vosk and Whisper (~150 ms on GPU), using OAuth PKCE with secure local token storage. Zero cloud, zero cost.
- EatTube — A Chrome extension that spins slots for a YouTube video to watch while you eat.
- ML-Library---Risk-Position-Sizing — A lightweight C11 ML toolkit for volatility forecasting and position sizing, built for time-series-safe evaluation.
Python · JavaScript · C · FastAPI · NumPy · Docker · AWS · Speech Recognition & NLP · Git
LinkedIn: https://www.linkedin.com/in/haegan-mcgarry-90aa43392/