AI Research Scientist | Research & Development

Westren Capital; Greater London; Hybrid; GBP 120,000 - 180,000; Full time

Job summary

Westren Capital, a London-based proprietary trading firm, is seeking a senior ML researcher to develop end-to-end models turning unstructured data into alpha signals. You will collaborate with traders to define constraints, design scalable architectures, and push models through real-world markets to endure latency and stability requirements.

Ideal candidates have 5+ years building ML systems, experience with transformers, HPC/distributed training, and publications or open-source contributions.

Qualifications

• Around 5+ years of experience building machine learning systems with real impact.
• Strong grounding in ML with exposure to transformers/LLMs.
• Proficient production-quality Python and/or C++, familiar with PyTorch/TF/JAX.
• Ability to communicate with trading researchers and maintain steady cadence.
• Experience with HPC or distributed training; GPU optimization a plus.

Responsibilities

• Spot problems that matter and push ML solutions through end-to-end development.
• Collaborate with traders to define constraints and translate data into signals.
• Develop models, tools, and infrastructure robust to real-world markets.

Skills

Machine Learning

Python

C++

PyTorch

Transformers

Tools

TensorFlow

JAX

CUDA

Job description

Westren Capital is a London-based proprietary trading firm specialising in digital asset markets. We are committed to rigorous, first-principles research at the intersection of quantitative finance and machine learning. We bring together talent from Mathematics, Physics, and Computer Science to push beyond conventional modelling approaches and translate cutting‑edge research into actionable signals across global markets.

Our culture is built around intellectual honesty, independence of thought, and a deep respect for evidence over narrative. We value researchers who are willing to challenge assumptions, explore uncomfortable ideas, and iterate quickly in the face of uncertainty. Collaboration is not ornamental here, it is structural. The best ideas tend to emerge where disciplines overlap and perspectives collide.

Our AI team sits at the core of this effort. It is a focused R&D group of quantitative researchers, engineers, and ML practitioners working on frontier problems in representation learning and large‑scale modelling. The mandate is simple in wording and difficult in execution: extract signal from unstructured data and convert it into robust, scalable alpha.

Responsibilities:

We are searching for researchers who have a history of using machine learning for solving challenging, realistic problems, not benchmark problems pretending to be progress. It's intrinsically end-to-end: spotting problems that matter, specifically where there's an advantage in developing LLM skills, and pushing them through the full development process.

You will be working with our traders, figuring out what constraints exist, what data looks like in practice, and what signals are realistically possible. From there, it becomes an exercise in iterating until your models, tools, and infrastructure don't collapse the first time you encounter real‑world markets.

The field is intentionally wide. Your projects could land anywhere on the research pipeline for quant finance, wherever you can turn unstructured data and cutting‑edge machine learning techniques into economically valuable insights.

We're not asking for excellence everywhere. The secret to success in this field lies in finding the right combination: expertise in one area, competence in another, and an innate interest in the third. Usually, this means machine learning, computer science, and some feel for how markets function when theory meets reality.And, inevitably, a few more duties nobody bothered mentioning but which will crop up regardless.

Requirements:

Around 5+ years of experience building machine learning systems that have delivered real, measurable impact, whether in industry or academia. A strong grounding in ML with some exposure to modern language models such as transformers or related architectures. Comfortable writing solid, production-quality code in Python and/or C++, and familiar with frameworks like PyTorch, TensorFlow, or JAX. Beyond tools, what matters is a mix of curiosity, range, and original thinking, balanced with a practical instinct for what actually works. You should be able to reason clearly about quantitative problems, communicate effectively with trading researchers, and maintain a consistent, dependable working rhythm.

Experience working with HPC environments or training large models in distributed settings, along with some exposure to GPU-level optimisation using CUDA or ROCm. A track record of taking models end-to-end, particularly in the context of LLMs, is valuable. Prior academic publications or meaningful contributions to open-source AI work are a plus. It also helps if you have considered views on how ML research and infrastructure should be done, and the judgement to know when those views need to bend to reality.