Case-study notes on product engineering, research systems, and the practical work behind shipping real software.
How GreekLab combines live data ingestion from four brokers, ML-driven regime detection, eleven strategy families, and a paper-trading engine to research Indian index options systematically.
How PatternRank turned years of market research into a US equities ranking product built around leakage-aware ML, daily refreshes, measured outcomes, and research-only positioning.
How VibeChess combined adaptive puzzles, real-time duels, Elo progression, bot fallback, subscriptions, analytics, and mobile launch flows into a shippable chess product.
How a Cosmos NFT marketplace went from one GitHub repo per chain to a single config-driven codebase — eliminating merge drift, diverging builds, and the overhead of shipping the same fix five times.