Three markets.
Four strategies.
One process.
We trade crypto, US equities and Hong Kong equities. Assets that once lived in separate systems now trade around the clock and increasingly price each other; we work where those markets meet.
- 01
Mid- and low-frequency strategies
Positions held from hours to weeks, built on market structure, flows and events rather than speed.
- 02
Directional systematic trading
Systematic strategies that take a view on direction — trend, momentum and regime — with sizing and exits defined in advance.
- 03
Liquidity provision
Providing two-sided liquidity where markets need it, managing inventory and risk across venues.
- 04
On-chain yields
Capturing yield opportunities on-chain with the same risk discipline we apply everywhere else.
Two desks share one risk framework and one execution stack. The discretionary desk is where traders form and act on their own judgment within clear mandates. The systematic desk runs as small pods, each led by a portfolio manager with their own strategies, researchers and execution.
An independent risk function sets and enforces limits across both desks.
Over $200 million in deployable proprietary capital. No outside investors, no redemption pressure.
Three principles sit behind everything above. They were in place before our first trade.
One process,
every market
Crypto, US and Hong Kong equities are covered by one research process, one risk framework and one execution stack. We look for opportunity in how markets relate to and misprice each other, not inside any one market alone.
Proprietary capital,
long horizon
We trade our own capital, with no external redemption pressure and no short-term distortion. That lets us validate strategies with real money, evaluate people on a reasonable timescale, and allocate capital to what has genuinely been proven.
AI-native
by design
From day one, Par Trading has run as three layers working together: people, systems and agents. AI accelerates research — unstructured data, factor discovery, hypothesis testing, backtest and analysis code — and runs much of the operating layer: trading-plan generation, execution monitoring, post-trade attribution. Researchers and traders focus on hypotheses, judgment and final confirmation.
Humans decide; agents operate.