Where the edge comes from
We build and run two families of systematic strategies: ones driven by machine-learned signals and delta-neutral statistical arbitrage. Different mechanisms, one method — hypotheses rigorously validated, positions sized to measured risk.
Forecasts you can trade on
Machine-learning predictions of short- and medium-horizon price moves, delivered as clean, versioned feeds. Validation runs in two stages: walk-forward while tuning, then an out-of-sample set for final sign-off. Feeds are monitored so they stay stable when regimes shift.
Directional forecasts
Triple-barrier predictions of price moves across horizons from seconds to days, with a confidence score you can filter entries or size positions against.
Trained on our data
We collect and store high-value market data, then rigorously engineer it into the features our models learn from.
Delivered as feeds
Models retrain automatically on fresh data and stream predictions as events, so your systems can adjust thresholds or sizing in real time. Not a report, not a notebook.
Earning from relative value, not direction
Market-neutral strategies that harvest structural spreads like basis, funding and cross-venue dislocations, while hedging out exposure to where the market goes.
Basis & funding
Capture perpetual funding and spot–futures basis while the directional risk is hedged away.
Cross-venue spreads
Exploit dislocations between venues and instruments, sized to how quickly they converge.
Neutral by construction
Positions are hedged by design, with explicit rules for risk, capital and net exposure.
Put them to work
Want these running for your desk or fund? Let's talk.