Quantitative Developer | HFT Crypto, Long-Range Dependencies & Attention Mechanisms | London

Joined
10/23/24
Messages
13
Points
3
A leading crypto-native proprietary trading firm in London is looking to hire a Quantitative Developer for one of its most successful teams. This group has a world-class track record and is responsible for a significant portion of the firm's PnL. They operate at the highest frequencies and are seeking to deepen their competitive edge by investing heavily in market microstructure research.

The core challenge of this role is to systematically data mine their vast reserves of market-making data to detect subtle anomalies and uncover hidden patterns. A unique aspect of their research is that they are targeting extremely long-range dependencies, with sequences potentially millions of data points long. This is not a simple time-series analysis task; they are looking for someone to design and build sophisticated models capable of capturing these complex dynamics. While they explore a range of techniques, including Hidden Markov Models, a strong understanding of modern sequence modeling, specifically attention mechanisms, is now a key requirement for this search.

This position requires a unique blend of skills:
  • Deep Statistical & Machine Learning Intuition: The ability to develop novel approaches for modeling complex, non-stationary data and capturing long-term memory in massive datasets.
  • Strong Development Skills: You will be responsible for the full lifecycle, from research and model implementation to production-quality code.
  • High-Frequency Experience: Demonstrable experience working with order book data in a systematic HFT context is critical. While direct professional crypto experience is ideal, they are open to individuals with a strong background in equities or FX market making, provided they have a genuine and deep-seated interest in crypto.
They are looking for someone with a degree from a top-tier university and a few years of relevant professional experience. This is an opportunity to have a direct and measurable impact on PnL within an elite team that is at the forefront of quantitative crypto trading.

Relevant skills:
  • Big'O + Trade-offs
  • Balanced Binary Search Trees (e.g., Red-Black Trees)
  • Skip Lists
  • Hash Maps + Doubly-Linked Lists
  • Vectors / Dynamic Arrays
  • Hidden Markov Models (HMMs)
  • Viterbi Algorithm
  • Baum-Welch Algorithm
  • Isolation Forests
  • Local Outlier Factor (LOF)
  • Clustering Algorithms (e.g., DBSCAN)
  • Log-Structured Merge-Trees (LSM-Trees)
  • Message Queues (e.g., Kafka)
Education preference:
  • Top 20 school: Maths and CS
  • TA for at least 1 module
  • High scores GPA
  • Coding & math challenges

Many applicants have strong mathematical modeling skills but lack the development expertise for high-frequency order books. The litmus test is whether they can code a Red-Black tree from scratch, unaided, in a timed interview.

The position is based in London.
 
Back
Top Bottom