02Technology

Proprietary intelligence, end to end.

Our edge is engineered. Artificial intelligence and machine learning sit at the core of a stack we design, build and operate ourselves.

The stack

Built in-house, owned end to end.

Control over data, models and execution lets us iterate quickly and trust what the system does under stress.

01

Artificial intelligence

Models fuse fundamentals, weather and market microstructure into actionable signals, continuously retrained as conditions shift. European power markets are shaped by interacting forces: weather drives renewable output, renewable output drives prices, and prices influence the behaviour of other participants. Our models are built to reason across these relationships, combining numerical weather predictions, generation and load fundamentals, and order-book signals. As market structure changes with seasons, new capacity and shifts in demand, the models are monitored and retrained so they reflect how the market behaves today, not how it behaved last year.

02

Machine learning

Forecasts of price, load and renewable generation are validated against out-of-sample data before capital is put at risk. Rather than producing single-point predictions, our models estimate probability distributions, capturing both the expected outcome and the uncertainty around it. Wind, solar, demand and price forecasts are treated as ranges to size against, not point estimates to trust blindly. Every model is validated on data it has not seen before it is considered for live trading.

03

Data infrastructure

A unified, low-latency data layer ingests exchange, fundamental and alternative sources into a single source of truth. Order books, transparency feeds, weather data, generation and load fundamentals, and alternative datasets are validated, time-stamped and reconciled in one consistent pipeline. This ensures research, production, trading and risk all work from the same numbers. It reduces silent data errors, keeps signals fresh, and allows new datasets or features to be added once and used across forecasting, strategy and risk.

04

Execution platform

In-house systems connect signal to venue, with deterministic risk checks on the critical path of every order. Owning the platform end to end gives us control over the full order journey, from model signal to market execution. Every order passes through automated pre-trade validation, including position, exposure and limit checks, and anything that would breach a limit is blocked before it can be sent. Because the guardrails are built into the execution path, the firm can act decisively when opportunities arise while maintaining strict control.

Pipeline

From raw data to executed trade.

01

Ingest

Market, fundamental & alternative data

02

Model

ML forecasting & signal generation

03

Decide

Strategy logic & risk constraints

04

Execute

Smart routing to venues

Künstliche Intelligenz
Maschinelles Lernen
Low latency
Cloud-native