"Independent AI Safety Researcher — Systems Architect of TQC — Governance and Policy and AI Forecasting"
I use AI as a precision instrument — directing architecture, making design decisions, and rigorously debugging until the system behaves exactly as specified.Every decision in my systems has a reason I can defend.
TQC — Three-Component Quantitative Trading System
- Engine: C++20 with lock-free MPSC ring buffer, thread-local buffers, zero shared mutable state on hot path
- Intelligence: HMM 3-state regime classification (Bear/Sideways/Bull), GARCH(1,1) volatility-scaled position sizing
- Analytics: kdb+/q data layer, Welford's online Sharpe/Sortino, walk-forward validation with stability verdict
- Infrastructure: Timing-safe authentication, SIMD math, Docker deployment, constant-time key comparison
The architecture mirrors live trading decisions exactly — not an approximation of them.
Where This Is Going This platform is designed to evolve. Near-term goals include establishing a systematic paper review pipeline covering the core AI safety and governance literature, a forecasting section with explicit resolution criteria, and deeper governance analysis of the international AI regulatory landscape.
The long-term goal is to contribute to governance frameworks that are technically grounded, institutionally realistic, and adequate to the risk profile of advanced AI systems. This requires both the technical analysis done here and engagement with policymakers — which is the next phase of this work.