Technical notes and essays on building AI tools, search interfaces, trading systems, and product engineering decisions.
Kalshi is a regulated prediction market: the edge is forecasting volatility, not price direction. This build note covers a LightGBM and PatchTST ensemble running on live Kraken WebSocket data and deployed with real capital.
The bottleneck for edge LLM workloads is usually the runtime, not just the model. air-runtime packages smart routing, speculative decoding, and KV-cache compression for constrained hardware.
Every LLM fails differently. orcha explores what happens when Codex, Claude, Gemini, and Kimi work in parallel and are forced into an explicit synthesis loop.
