Bayes Strategies
Bayes Strategies provides an event-driven optimization engine that continuously re-dispatches customer batteries, solar panels, and flexible loads over a 48-hour rolling horizon. The service re-runs a constrained cost-minimization problem whenever asset states or market forecasts update, emitting live steering signals without manual intervention. The platform's data backbone is orchestrated on Dagster, supporting backtesting and feature-computation pipelines.
- Energy
- Clean Technology
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
Energy asset owners and operators with batteries, solar panels, and flexible loads face the challenge of optimizing dispatch decisions in real time. Market prices and asset conditions change continuously, making static scheduling inefficient and leaving potential cost savings on the table.
Solution
Companion.energy provides an event-driven optimization engine that continuously re-optimizes how customer energy assets are dispatched over a 48-hour rolling horizon. Whenever asset states or market forecasts update, the engine re-runs a constrained cost-minimization problem and emits live steering signals. The service is built on a ground-up redesign of the platform's core data model, codebase, and backtesting pipelines, orchestrated on Dagster, to ensure reliable real-time decision-making. This allows customers to automatically respond to market dynamics without manual intervention, reducing energy costs through optimized battery charging, solar usage, and flexible load shifting.
Target Audience
Primary customers are commercial and industrial energy asset operators managing batteries, solar installations, or flexible loads who need automated, cost-optimizing dispatch decisions in volatile energy markets.
Features
- Event-driven re-optimization over a 48-hour horizon triggered by any update to asset states or market forecasts
- Constrained cost-minimization solver that produces live dispatch steering signals for batteries, solar, and flexible loads
- Dagster-orchestrated data backbone including backtesting and feature-computation pipelines that underpin steering, forecasting, and reporting
- Continuous operation as a real-time backend service, not a batch process