Research Technology Developer

The team

The Research Technology team at XTX Markets takes care of all aspects of our research infrastructure. Tasks range from writing the software which manages fair and efficient distribution of work on our compute cluster to tuning operating systems, storage, and networks, evaluating high-end leading edge new compute technology (often before it’s formally released) and designing the datacentre environment for it to run in. We are a full stack team that works side-by-side with our quantitative researchers to make the most performant, reliable and transparent system we can on one of the larger HPC clusters in existence.

The role

The right candidate will:

- Contribute to all components of our HPC infrastructure and code, and work on growing one of the biggest private compute clusters anywhere.
- Write software that runs on a compute cluster that grows continually but currently has ~1PB memory, ~100K CPU cores, 1000’s of compute offload devices, 60+PB of high performance storage, connected by a multi-Tb networks.
- Enter an environment where improvements can usually be made very quickly, and where the results of those changes are both immediately visible and can make a large impact to the quantitative research function at the heart of our business.

The skills

Strong coding skills, preferably with recent exposure to python and at least one statically typed language.
- A working knowledge and history of:
o Delivering features in large-scale distributed systems
o Implementing software workflows
o Testing and deployment methodologies
- A desire to solve complex problems optimally from the ground up, not just always reassembling components written by others.
- Using your knowledge from many layers of the technology stack (network/hardware/os/software) to produce an optimal result.

We would not expect someone to have all the following skills or experience, but some subset would be preferred and gives an idea of the wide range of technologies we work with:
Prior experience with machine learning frameworks like pytorch or supporting technologies like CUDA.

- Understanding of large-scale batch processing systems; approaches to efficient scheduling; and dependency management
- Working with compute offload devices eg: GPU’s (but also some other more niche hardware)
- Experience with computer networks
- Experience with large-scale parallel storage

The candidate

- Will have previous direct experience in a similar setting with large-scale-compute, although previous experience in financial services is not necessary.
- Proven track record of delivering complex projects with a 1 day to 1 year horizon, with 4 or more years of relevant experience.
- A good STEM degree from a reputable tertiary educational establishment is preferred.
- Top-notch technical credentials, and a drive to achieve is essential.