Candle Continuity, Local-First Market Data, and Deploy Health
A 32-commit day making candle-service fast and continuous while moving dex-trader to local-first data.
A 32-commit day making candle-service fast and continuous while moving dex-trader to local-first data.
Docs in the image, write serialization for tick ingest, dynamic pricing scaffolding, and a Solana DEX backfiller.
Stopping paper trades on stale data, uncapping freshness so pools aren't starved, and broadening the source pool.
Bringing two dex-trader branches together phase by phase while overhauling roles and wiring a daily training-data pipeline.
How I used LLMs and ARQ to build a self-adapting, scalable web scraper that survives real-world site changes.
I rebuilt Git Context’s analysis pipeline to handle large repos with batching, caching, and smarter dependency management.
I refactored HomeForged's schema analysis system to unify entity handling—here's how pure analyzers and canonical structures changed the game.