Compute Researcher

at  GResearch

Do you want to tackle the biggest questions in finance with near infinite compute power at your fingertips?
G-Research is a leading quantitative research and technology firm, with offices in London and Dallas. We are proud to employ some of the best people in their field and to nurture their talent in a dynamic, flexible and highly stimulating culture where world-beating ideas are cultivated and rewarded.
This role is based in our new Soho Place office – opened in 2023 - in the heart of Central London and home to our Research Lab.

The Calc Farm Platform Services team develops and manages a large calc farm platform that enables the business to conduct complex research at scale.
The calc farm platform is absolutely critical to our business and plays an integral role in how we conduct research. We are seeking a highly motivated individual to join the team to find and propose solutions to help us realise the full potential of our platform.
Modern research techniques are becoming ever more complex with various ML and data science frameworks, DAG schedulers, job schedulers, container orchestrators, and physical hardware all needing to work together in harmony.
Owning and controlling the whole stack provides us with an extraordinary opportunity to optimise it, which is why we’re seeking an individual to focus on delivering gains over multiple time horizons.

Key responsibilities of the role include:

• Understanding the technical aspects of short and long-term customer needs and researching ways in which those needs can be met, including both novel and industry-standard solutions
• Proposing and building consensus around long-term solutions
• Working with engineers and customers to incrementally build towards those solutions

This is a deeply technical role using quantitative methods to produce analysis and, more generally, build an environment where other individuals, such as architects, can make the right choices.
The initial focus will be on Armada, an exciting open source CNCF project, built and maintained by the team, which we use to solve multi-cluster Kubernetes batch job scheduling at scale.