Where we work
We focus on the parts of a trading stack where speed, correctness and control all have to hold at once.
Low-latency infrastructure
Latency is rarely one problem. It is a long tail of small delays and sources of jitter across the network, the host and the application. We find them, remove them, and make sure they stay gone.
- Where time goes. Profiling from the wire to the application, using hardware timestamps and packet captures.
- Hosts and operating systems. CPU isolation, interrupt affinity, memory and NUMA layout, kernel and NIC configuration.
- Network paths. Switching, multicast and connectivity for market data and order flow.
- Staying fast. Latency monitoring with percentiles and alerts on drift, so regressions are caught before they matter.
Order management & order books
The order path is where correctness matters most. We build order and market-data systems that behave the same way every time, recover cleanly, and leave a complete record.
- Order lifecycle. New, amend, cancel and fill handling with deterministic state and clean recovery.
- Routing and gateways. Exchange connectivity and protocol handling, including FIX and native binary protocols.
- Market data and books. Feed handling, normalisation and order-book building.
- Audit trails. Complete, queryable records of every order event.
Risk systems & controls
Controls only work if traders and operators trust them and they never slow the business down. We put risk checks in the order path and make them fast, clear and dependable.
- Pre-trade checks. Limits, price bands, order-size and credit checks.
- Real-time exposure. Positions and limits updated as orders and fills arrive.
- Kill switches. Controls operators can reach quickly and rely on.
- In the path, not in the way. Checks designed and tested for minimal added latency.
AI in engineering
We use AI tools to make engineering work faster and more thorough, under the same discipline as everything else: measured, reviewed and owned by people.
- Testing. Generating test cases and analysing replay results.
- Telemetry and logs. Surfacing anomalies and patterns in large volumes of operational data.
- Research tooling. Internal tools that help engineering teams explore data and ideas faster.
Tell us about the system.
Describe what you're building or what's holding it back, and we'll get back to you.