I’m an independent builder—just shipped a general-purpose, recursive agent kernel, designed for risk-managed, auditable decision-making in any regime-driven environment (trading, robotics, anomaly detection, etc).

Theory Paper & Artifact:

Zenodo (600+ downloads in under a month): https://zenodo.org/records/15713252 — Full theory, architecture, and open artifact — PDF, code skeleton, results, and methodology

Latest live demo/report:

LinkedIn (clean, out-of-sample test vs. SPY, posted July 16): https://www.linkedin.com/posts/basil-iftikhar-34399318a_recu... — Transparent, auditable, with pilotable code and trade logs — Sharpe 1.12, Max Drawdown -3.97%, competitive returns, full exec summary/report

Happy to answer any technical or practical questions—feedback and critique welcome!