Production systems running on self-hosted infrastructure. Quantitative analysis, automated trading, and macro-economic intelligence — built from scratch, operated 24/7.
A distributed quantitative trading system. Real-time cryptocurrency data ingestion, technical analysis, signal generation, and automated paper trading — processing hundreds of symbols simultaneously on a bare-metal Kubernetes cluster.
431 USDT perpetuals ingested across 5 timeframes (1m to 1d). Tiered symbol selection by volume. Idempotent batch processing for fault tolerance.
RSI mean-reversion strategy with multi-indicator confirmation. Market regime detection filters counter-trend entries. Dynamic stop-loss and take-profit calculation.
Position limits, per-trade risk caps, and a daily loss circuit breaker. Automated paper execution with real-time P&L tracking and alerting.
A large-scale data analytics engine that ingests billions of data points and runs quantitative analysis to uncover statistical relationships across crypto markets and macroeconomic signals.
Cointegration (Engle-Granger), spectral analysis (Random Matrix Theory), PCA, network analysis with community detection, and regime identification.
OHLCV data from July 2017 to present across 6 timeframes. 73 macro indicators from FRED, Yahoo Finance, ECB, and sentiment sources.
Automated via Kubernetes CronJobs. Network and spectral analysis run hourly. Data ingestion scheduled across daily, weekly, and monthly cycles.
A multi-source sentiment engine that collects macro-economic and social signals, computes a unified fear/greed index, and classifies market regimes — from extreme fear to extreme greed.
VIX, DXY, credit spreads and yield curves from Yahoo Finance & FRED. GDELT global news sentiment. Google Trends across 13 economic keywords. YouTube trending sentiment across 5 countries.
Weighted index: VIX (25%), credit spreads (15%), news tone (15%), policy uncertainty (15%), yield curve (10%), social sentiment (10%), Google Trends (10%). Regime labels from extreme_fear to extreme_greed.
Dual-write pattern (Kafka + direct DB), automated compression and retention policies, full Ansible deployment automation.