AI-native trading ยท Hong Kong
The next-generation
trading firm.
A small team. Thousands of agents. A global trading firm.
A few exceptional people build the system. Agents research, experiment and trade across global markets.
Operating model
Built to scale through agents
One shared system expands research, experimentation and market coverage, so each new market starts from everything learned in the last.
Small human team
Builds and governs the system. Humans set direction and approve what goes live.
Massively parallel agents
Run research, software and real-hardware experiments side by side, across markets.
Selected strategies
Only the strongest candidates advance into a monitored live portfolio.
Shared learning
Validated improvements to methods and infrastructure feed into the next experiment.
Our advantage
Built to learn at scale
Thousands of experiments, and one system that learns across all of them. We have been built around agents from day one.
Parallel research
Agents pursue many candidates across markets at once. The strongest advance to trading.
Shared learning
Validated improvements in methods and infrastructure carry into the next experiment.
Better foundation models
New model capabilities enter the same system and, once evaluated, scale across every workflow.
Research
Research in hours, reused across markets
Agents take raw market data to tested strategies, and the infrastructure built for one market becomes the starting point for the next. We also study the research process itself: agents that improve their own methods.
Our research →Hardware
Parallel experiments on real hardware
Agents automate design, deployment and testing on real FPGA boards. Our model-to-RTL toolchain turns a strategy into measured hardware in a fraction of the usual cycle time.
Hardware research →From the blog
Latest writing
Work with us
We are a small team, and we hire people who build systems that let agents do the work.