Project Overview

The Quantitative Theory Engine is a personal research project focused on learning quantitative stock analysis and developing automated investment strategies. It is decoupled from my professional information security roles and serves as a technical sandbox for algorithmic finance.

Infrastructure & Stack

The engine is hosted within my local Debian VM Lab environment on a Proxmox hypervisor.

Host Hardware: 96GB RAM dedicated to intensive modeling tasks.
Orchestration: Docker containers manage the core processing units.
AI Integration: Utilizing Ollama (Llama 3.3 70B & Qwen3-Coder) for natural language data processing and code optimization.

Proxmox Lab Architecture for Quant Modeling

Technical Implementation

The system utilizes a modular architecture to separate data ingestion, strategy backtesting, and execution logic. By leveraging the local Lab’s high-memory capacity, the engine can process historical ticker data without external cloud compute dependencies.

Current Progress:

Static Site Documentation: Migrating engine logs and design docs from WordPress to this Hugo-based Markdown system.
Local LLM Agents: Implementing local AI frameworks to assist in identifying anomalies in financial datasets.
Sample Backtesting Output - Feb 2026: The results of quantitative stock analysis, specifically backtesting, performed.

Sample Backtesting Output - Feb 2026

[!NOTE] This project is for personal research, it remains separate from my career and professional activities.