The short version

We run an HFT trading stack across global crypto markets. One of the things that stack spits out is market data, order books, trades, derivatives, the lot... tured clean and fast because we need it that way to trade. Turns out it's good enough that other people pay us for it, and now it's a product in its own right: one of the best crypto market data services anywhere.

We need a data engineer to help us run it like a product instead of a happy accident.

The Role

We are in search of a Quantitative Researcher possessing exceptional analytical capabilities, as demonstrated by an advanced degree in mathematics, statistics, physics, computer science, engineering, or similar fields. A profound enthusiasm for systematic research and market microstructure is required.

Prior experience in high frequency trading is expected and candidates who exhibit an entrepreneurial spirit and the belief that any problem, no matter its complexity, can be addressed, are highly regarded.

Preference is given to individuals who are either located in, or open to self-relocating to London, valuing the benefits of an in-person work research environment among peers with similar interests.

This position offers a distinct chance to play a pivotal role in the evolution of our core trading engine and market making strategy. You will collaborate intimately with the founding team on vital and intricate aspects of our operations, assuming responsibility for specific areas from the outset. The role affords ample opportunity for proactive engagement in shaping your duties, contributing to meaningful and enduring advancements within the company.

What you will work on

You'll own the pipelines that take raw exchange feeds and turn them into something clients trust their money to.

• Build and maintain ingestion from CEX and DEX venues — WebSocket feeds, REST, on-chain — and keep them alive when exchanges do exchange things.

• Normalise, validate and reconcile the data. Gap detection, dedup, schema drift, the unglamorous stuff that's the entire game in market data.

• Own storage and delivery: time-series and historical datasets, the APIs and feeds clients pull from, and the SLAs behind them.

• Build the monitoring and alerting so we know data quality has slipped before a client does.

• Keep the hot paths fast. Some of this lives in Rust for a reason.

Requirements

• 2–4 years building real data infrastructure that other people depended on. Not coursework, not a notebook that ran once.
  - Strong Python, strong Rust. Polars in particular — we don't reach for pandas here.

• Comfortable on AWS and on the command line. You can stand up your own infra and debug a distributed system without panicking.

• Solid SQL and a real instinct for data quality — you notice when numbers are subtly wrong, not just when a job fails loudly.

• Sharp, curious, low ego, allergic to corporate posturing. You'll be close to the people who depend on your work, so you'll hear about it fast either way.

Nice to have

Any of these moves you up the list, none are dealbreakers:

• Crypto, trading, or market microstructure exposure. If you know what an order book is and why a tick matters, say so.

• Exchange APIs and real-time/streaming systems (Kafka, Redpanda, NATS).

• Time-series stores (ClickHouse, kdb+) and the Arrow/Parquet world.

• Infrastructure-as-code (Terraform) and CI/CD.

How we work

Small team, no layers, no theatre. You'll own things end to end and your work will be in front of paying clients quickly. We're in the office five days a week in London because the best version of this job happens in the same room, fast.

If the data's wrong, it's our problem before it's the client's. That standard is the job.

Location

Based in our London office