Inside voltage-kalshi: Building a Kalshi BTC Volatility Bot
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.
voltage-kalshi is a live volatility stack for binary BTC contracts on Kalshi. The system watches market data, builds short-horizon features, and turns model outputs into an execution loop designed for real conditions rather than clean retrospective charts.
The interesting constraint is not only the model. It is the operating surface around the model: streaming ingestion, order-book signal design, risk-aware execution, and enough visibility to understand why the system wants to trade.
Why volatility, not direction
Kalshi BTC contracts reward a different kind of prediction than spot trading. The question is whether a range or threshold resolves in a given way, so useful signal often comes from volatility, structure, and timing rather than a simple bullish or bearish call.
That changes the system design. The bot needs fresh market state, features that reflect short-horizon movement, and a way to decide when a forecast is strong enough to justify execution risk.
System shape
The stack ingests Kraken WebSocket data, builds features around recent BTC behavior, and combines gradient-boosted and sequence-model style signals. The important engineering surface is the loop around the model: how data freshness, order-book context, and risk controls shape the final action.
The portfolio presents voltage-kalshi as proof that I can move past notebook-only modeling into live systems where latency, capital, and operational visibility matter.
What I would improve next
The next useful iteration would separate research backtests, paper trading, and live execution into clearer lanes with reproducible metrics. I would also make the operator dashboard more explicit about confidence, exposure, and why the system decided to stay out.
Related work
voltage-kalshi
A live Kalshi volatility stack built around streaming BTC data, feature engineering, and model-driven execution with real capital on the line.
Core challenge: Pricing short-horizon volatility under noisy live conditions instead of relying on retrospective backtests alone.
- Kraken WebSocket ingestion
- Kalshi order-book signal design
- LightGBM plus PatchTST execution loop