Skip to content
BAR.
← All work
Private2025 — 2026

diBoaS Analytics

The quant brain: backtesting, Monte Carlo risk, ML anomaly detection.

diboas.com/market, the live macro-regime dashboard fed by the private engine
diboas.com/market, the live macro-regime dashboard fed by the private engine
Role
Sole engineer: architecture, quant methodology, data pipelines
Stack
Python · pandas · scikit-learn · scipy · 808 tests · 19 CLI commands

A ~50,000-line Python engine that is the quantitative backend of diBoaS. It collects data from five-plus financial sources behind circuit breakers and rate limits, pushes everything through a four-gate quality pipeline, backtests ten risk-tiered strategies, and runs regime-switching Monte Carlo simulations. Markov transitions between market regimes, Student-t fat tails, VaR and CVaR, plus an Isolation-Forest anomaly layer and real-time protocol health monitoring with alerting.

The public window into it is diboas.com/market: a live macro-regime dashboard fed by a build-time pipeline with dual-source verification and a fail-closed quality gate.

What it proves

  • Institutional-grade quantitative methodology, not toy finance code
  • Data engineering as a discipline: quality gates, freshness SLAs, audit trails, reproducible runs
  • Applied ML where it earns its keep: anomaly detection, not decoration