Systematic 0DTE Options Trading
Defined-risk options strategies on SPX and index products, driven by dealer positioning and regime. These systems carry strong Calmar and Sortino ratios across forward testing and live management.
Algorithmic Trading · Real-Time Financial-AI Systems
Real-Time Market-Structure Intelligence
Real-time 0DTE SPX options-microstructure infrastructure, with a self-hosted AI layer that reasons over live market structure. Built and operated for research purposes only. Nothing here is investment advice, a recommendation, or a solicitation.
The options analytics behind this work - full chain, historical and live, with millisecond-accurate real-time flow - come from QuantData, my primary data source. 20% off for readers: quantdata.us/?discount=KGB.
Some methodologies
Focus areas
Real-Time Market Microstructure
Tick-level ingestion and per-minute aggregation of gamma / delta / vega exposure, charm, skew, max-pain and net-premium drift across the full 0DTE SPX chain, VIX and 250+ concurrent instruments.
Applied AI for Trading
Local LLM agents that learn the way I trade — from my notes, daily write-ups and past sessions — through a pgvector memory layer, pre-warmed with dense structural summaries instead of raw ticks.
Quantitative Research & Backtesting
A strategy registry with walk-forward validation and prop-firm drawdown/survival modeling, with reproducible result caching for honest, comparable evaluation.
What I'm currently building
Self-hosted inference for real-time trading
Today I rotate one-shot calls across commercial APIs — Claude (the most capable), plus Gemini, Mistral and others — feeding each a tightly compressed market summary, because tick-level data won't fit and the plan caps run out mid-week. Moving to a self-hosted model removes the cap and lets me benchmark what actually matters for trading:
Latency
Time-to-first-token and generation latency low enough to fit inside scalping execution windows.
Requests per minute
Sustained, minute-by-minute context injection through the session without thermal or VRAM bottlenecks.
Data throughput
Rapidly experimenting to learn which real-time signals carry edge — and which are noise — across high volumes of structured market data.
Execution integration
Decision speed fast enough for the model to react inside scalping windows — quicker than manual execution of the same methodology.
The longer-term goal is a refined local model with durable long-term memory — warmed from summary profiles and continuously taught my methodology — running at high throughput on dedicated hardware.
Selected work
Systems I've built
SIGNUM-ONE Options Intelligence
A real-time SPX options analytics and automated-reporting platform: per-minute dealer-exposure snapshots, filtered net-premium drift, regime detection, and gated end-of-day / pre-open / intraday report generation — surfaced through a React charting portal, with verification gates on every published figure.
Local AI Agent Stack
Multi-agent orchestration over a knowledge base of thousands of embedded research chunks with structured tool use. It learns from my daily write-ups and past sessions via pgvector — and is being moved off API caps onto self-hosted hardware at near-zero marginal cost per token.
Research & Backtest Engine
Backtesting, a strategy registry and signal research (dealer-positioning, regime and ML families) with prop-firm survival simulation and reproducible, cached result sets for comparable evaluation.
PKI, WebTrust & Cryptography
Earlier career · 2001–2015
Two generations of WebTrust-certified public-key infrastructure — multi-layer CAs, registration authorities, LDAP, and OCSP — deployed to government, banking, and industrial clients across 30+ countries. HSM-integrated custom cryptography, smart-card platforms, and a granted patent for digital authentication of valuable goods. Secure data centres built to WebTrust and DoD Level 5 specifications; annual WebTrust audit programmes; EU EEMA PKI interoperability standards work.
Digital identity, social networks & blockchain
Earlier career · 2001–2015
Secure credential and password-management platforms across web, desktop, and mobile — X.509 digital IDs, encrypted local stores, and biometric integration. Consumer social-network systems for major sports franchises with Facebook/Twitter APIs, content management, and fan push alerts. Smart-card authentication architectures; commercial certificate and subscription transaction portals. Cryptocurrency and blockchain identity infrastructure for digital-asset provenance and secure transactions.
Daily intelligence
End-of-day reports
Every session my stack generates an end-of-day market report — dealer-positioning, regime, flow and the levels that matter — fronted by $RAVOLM, a regime-aware animated volume visualisation. A few recent ones:
2026-10-19 · Monday
OPENED 7,702.77, MADE THE 7,684.79 LOW AT 10:04 AND THE 7,731.18 HIGH AT 11:37, AND CLOSED 7,690.73 (+0.41%)
SPX +0.41% | open 7,702.77, low 7,684.79 at 10:04, high 7,731.18 at 11:37, close 7,690.73
2026-10-15 · Thursday
OPENED 7,687.31, MADE THE 7,669.66 LOW AT 09:57, ROSE TO A 7,754.15 HIGH AT 15:23 AND CLOSED 7,744.06 (+0.33%)
SPX +0.33% | open 7,687.31, low 7,669.66 at 09:57, high 7,754.15 at 15:23, close 7,744.06
2026-10-14 · Wednesday
OPENED 7,774.18, MADE THE 7,789.15 HIGH AT 09:49, FELL TO A 7,693.31 LOW AT 14:01 AND CLOSED 7,718.94 (-1.11%)
SPX -1.11% | open 7,774.18, high 7,789.15 at 09:49, low 7,693.31 at 14:01, close 7,718.94
2026-10-13 · Tuesday
OPENED 7,795.27, MADE THE 7,814.15 HIGH AT 09:39, FELL TO A 7,786.11 LOW AT 11:38 AND CLOSED 7,805.45 (-0.02%)
SPX -0.02% | open 7,795.27, high 7,814.15 at 09:39, low 7,786.11 at 11:38, close 7,805.45
2026-10-09 · Friday
OPENED 21PT HIGHER, DIPPED TO 7,779.34 BY 09:48, THEN CLIMBED ALL DAY TO 7,820.57 AT THE CALL WALL, +0.59%
SPX +0.59% | open 7,786.36, low 7,779.34 at 09:48, high 7,820.57 at 14:51, close 7,811.54
2026-10-08 · Thursday
SPIKED TO 7,797.79 AT 12:17, THEN FELL TO 7,731.26 BY 13:28, MOSTLY IN NEGATIVE GAMMA, BEFORE RECOVERING HALF, -0.47%
SPX -0.47% | open 7,778.45, high 7,797.79 at 12:17, low 7,731.26 at 13:28, close 7,765.36
2026-10-07 · Wednesday
GAPPED DOWN 26PT, FELL TO 7,763.34 BY 10:53, THEN CLIMBED BACK TO 7,807.02 AS NEAR-SPOT GAMMA TURNED POSITIVE, -0.22%
SPX -0.22% | open 7,792.98, low 7,763.34 at 10:53, high 7,807.02 at 15:02, close 7,801.77
$RAVOLM (Regime-Aware Animated VOLM) — research/visualisation, not financial advice.
Curriculum vitae
Profile
CISO and CTO background with 20+ years in regulated financial services and digital security — alongside full-stack systems development at the intersection of quantitative trading and applied AI. Designs, builds and operates systems such as real-time 0DTE SPX options analytics, dealer-exposure and regime engines, and a multi-agent AI / reporting layer with pgvector memory. These trading systems carry strong Calmar and Sortino ratios across forward testing and live management — with work now underway toward self-hosted LLM inference to apply local models to trading methodology at high throughput.
Core competencies
- Systematic options trading — defined-risk SPX / 0DTE strategies; risk-adjusted performance (Calmar, Sortino) across forward testing and live management
- Real-time market-data engineering — DXLink / WebSocket ingestion, TimescaleDB hypertables, Redis Streams event bus, Dockerised microservices on Linux with systemd
- Options analytics — GEX / DEX / VEX, charm, dealer positioning, max-pain, skew & term structure
- Quantitative research & ML — Backtrader walk-forward validation, prop-firm survival modeling, scikit-learn / XGBoost signal work; Parquet-derived cache layers for bulk pandas replay (Postgres as source of truth, columnar extracts where SQL would be the wrong tool)
- Live trading integration — broker and execution APIs across Tastytrade, Schwab (thinkorswim), MooMoo (Futu OpenD API), and cTrader (prop-firm FIX / Open API + Protobuf); DXLink market-data feeds
- Full-stack web engineering — FastAPI backends, React + Vite + Tailwind front ends, Plotly interactive charting and research portals
- Applied AI & agent systems — multi-agent orchestration, RAG over pgvector and Neo4j, tiered LLM dispatch with rate-limit metering, structured tool use, self-hosted inference (vLLM / Ollama)