AI systems · Quant R&D · autonomous research workflows
Building an agentic Quant R&D system for trading strategy research.
Alongside native product work, I design and operate advanced AI-assisted environments for quantitative trading systems. They combine software engineering discipline, MT5/MQL5 automation, robustness testing and autonomous agent workflows to move from hypothesis to implementation, experiment, validation and research notes.
One system operates as a tool-using Quant R&D agent working directly with code, terminal workflows, backtest artifacts and research logs. Another reads Telegram trading-signal channels, interprets noisy real-time messages through LLM models and routes structured decisions into live trading-platform execution workflows.
My public quant research and strategy portfolio is available at deusquant.com.
Quantitative TradingAI AgentsLLM Signal AnalysisTelegram AutomationRAGMQL5 / MT5PythonReal-time ExecutionBacktestingWFO / OOSRobustness Testing
Research → implementationTranslate trading hypotheses, papers and strategy ideas into testable EA variants and reproducible experiments.
Automation → artifactsCompile, run, collect and organize MT5/MQL5 backtests, optimization outputs, CSV reports and research notes.
Real-time signal automationIngest Telegram trading-signal channels, normalize unstructured messages with LLMs and trigger structured execution flows on a trading platform.
Validation → disciplineTreat profit factor and in-sample performance as research filters — then apply OOS, tick validation, WFO and robustness checks.
Multi-agent engineeringUse planning, debugging, code review and delegated subagents to inspect code, design tests, analyze results and reduce implementation risk.