OUR APPROACH
Research sets the direction. AI and automation are built into how we investigate ideas, develop models, and turn evidence into a repeatable trading process. Our focus is on volatility risk premia and commodity futures term structure.
We start with a question about why a risk premium exists and who bears the risk. In volatility, that means understanding the demand for protection. In commodities, it means examining storage, inventories, carry costs, and commercial hedging needs. These economic relationships give our research a basis for interpreting the data.
Our approach is AI-first: we use it to support the research workflow, from organising information and developing analytical code to exploring hypotheses and examining results. The purpose is to shorten the distance between a question and a useful experiment. Models and outputs remain subject to the same standards of evidence, reproducibility, and review as the rest of our work.
We examine data quality, sensitivity to assumptions, and performance across different market conditions, including periods outside the original research sample. Trading costs, contract rolls, liquidity, and margin requirements belong in that assessment. A promising backtest needs an economic explanation and results that withstand further testing.
Automation connects data preparation, strategy evaluation, portfolio construction, and trading workflows. We aim to make recurring decisions consistent, research reproducible, and changes traceable. Exposure limits, position sizing, and operational checks are part of the implementation, alongside the logic that generates a trading signal.
Live behaviour informs the next round of research. We review exposures, implementation costs, and differences between expected and observed outcomes. Automated monitoring supports that feedback loop, while research and risk judgement guide decisions about whether a strategy should be maintained, adjusted, or reassessed.