Quant Researcher

The role

Our extended fields are derived features computed from full-depth data: multi-level order flow imbalance, VPIN, queue dynamics, cancel-to-fill ratios, book resiliency, price impact, markouts and participant flow. Each one is computed on our capture clock and published on a 100 ms grid. Customers will use them as model inputs, so every field has to be defined precisely, computed the same way on every venue and tested before it ships.

Our first study found that adding these fields to a baseline of raw trades and top-of-book data lifted the out-of-sample directional hit rate on Hyperliquid BTC, ETH and SOL perpetuals from 51.6% to 54.9%, with the same model and hyperparameters. You will extend that work to more fields, more venues and more horizons, and you will be the person who decides when a result is good enough to publish.

What you will do

• Define new extended fields and refine the existing eight, writing the specification that engineering implements and the documentation customers read.
• Build and own the evaluation framework: walk-forward splits with no overlap, fixed models and hyperparameters across feature sets, and reporting that states sample, horizon and limits.
• Run the studies we publish on our research page, from design through write-up, including the ones that find nothing.
• Check the data from a researcher's side: spot timestamp anomalies, book states that should not exist and fields that behave differently across venues.
• Work with Maya on how L3 and L4 events feed the queue and participant-flow fields, and with Lukas on how fields are delivered in the live stream.
• Join partner calls when a customer's question is about methodology.

What we are looking for

• Three or more years of quantitative research at a trading firm, fund or market maker, or a PhD with empirical work on high-frequency or microstructure data.
• Fluent Python for research (NumPy, pandas or Polars) and comfort with tens of millions of rows per day per instrument.
• A solid grounding in statistics and time-series methods, and practical habits against overfitting: out-of-sample discipline, multiple-testing corrections and honest baselines.
• A working understanding of limit order books, queue priority and what moves short-horizon prices.
• Writing that a skeptical reader can follow and check.

Nice to have

• Experience with crypto perpetuals, funding rates or on-chain data.
• Published research in market microstructure, or public notebooks and write-ups we can read.
• Some Rust or C++, enough to read the code that computes a field.
• Experience presenting research to traders or portfolio managers.

Compensation

Base salary of $150,000 to $185,000 a year and stock options for 0.25% to 0.50% of the company, vesting over four years with a one-year cliff and a ten-year exercise window after you leave. The band is the same wherever you live, paid in USD or local currency, directly or through an employer of record.

The role includes 25 days of paid leave plus public holidays, health insurance, 16 weeks of paid parental leave, a $3,000 equipment budget, a $2,000 yearly learning budget and an annual company week in person. Research conference travel comes out of a separate budget. See the careers page for the full list.

Location and hours

Remote, anywhere from UTC−5 to UTC+1, which covers the US East Coast, the UK, Ireland, Portugal and western Europe. You will work most closely with Daniel in New York. Everyone keeps 13:00 to 15:00 UTC free for live discussion. There is no on-call duty for this role.

Process

1. Application, read by a founder, with a reply within a week.
2. 30-minute intro call with Chris.
3. 90-minute research conversation with Daniel: you design a study on a question we give you a day in advance.
4. Paid work sample, up to four hours, at a flat $400: define one feature on a day of Hyperliquid L2 data and evaluate it against a baseline.
5. Calls with Maya, Lukas and Priya, plus two references. Decision within three business days.