T+00Choose
Act under partial information.
The model sees an incomplete state and commits before the full consequences are observable.
Reinforcement learning environments
We build market-derived RL environments to teach agents applied ML and long-horizon planning under adversarial noise. Static synthetic benchmarks are unrealistic and saturate quickly, whereas markets are non-saturating and self-improving.
THESIS
Static worlds produce static intelligence
Quant is the hardest, yet solveable data science task.
We build environments from real market data and test agents on different quant researcher task inside it. Agents receive (some) professional quant researcher tools and bash to make its own tools, and have to make trading decisions in our environments to produce profitable strategies.
We use markets because they are self-improving and do not saturate: better trading makes markets more efficient, which makes better trading harder again. Strategies that worked in the past do not always work in the present because other participants discover them and trade the edge away, or because the market regime changed. All these properties make our environments the ultimate benchmark to evaluate ever improving models.
HORIZON
A decision is not a moment
Trading decisions affect more than one future step.
Trading successfully means planning ahead multiple steps and assess trade-offs between short and longterm gains
T+18HCompound
The decision becomes part of the environment.
Exposure, opportunity cost and every action not taken reshape the path that follows.
T+53HRevalue
The objective moves.
A decision can remain locally correct while becoming globally expensive as conditions drift.
T+96HAdapt
The policy that worked has expired.
Success belongs to the model that recognizes the new regime before yesterday’s behavior becomes consensus.