This ML Systems Engineer role focuses on training and infrastructure for software-engineering language models. It suits a hands-on engineer who can work across machine learning, production software, data pipelines and distributed infrastructure.
At a glance
- London-based with full-time office working as the default
- £80,000–£110,000 base salary plus £80,000–£110,000 in share options
- Start date is as soon as possible
- Typically suited to candidates with 3–5 years of experience or equivalent ability
The role
- Build systems and environments used to train models that write, test and maintain software
- Work closely with machine learning researchers, infrastructure engineers and product teams
- Take ownership of projects from initial idea through deployment and iteration
What you will do
- Implement supervised fine-tuning and reinforcement-learning pipelines
- Develop custom PyTorch data loaders, training objectives and evaluation workflows
- Create, filter, transform and sample large-scale training datasets
- Build containerised training and evaluation services using Docker and orchestration platforms such as Kubernetes
- Design evaluations covering tests, benchmarks, repository tasks and engineering workflows
- Analyse model failures, regressions and reward-hacking risks
What you will bring
- Strong software-engineering fundamentals and experience with non-trivial production code
- Ability to work primarily in Python and Go
- A focus on correctness, maintainability and code quality
- Confidence handling ambiguous technical problems independently
Working pattern and location
- Based in London
- Full in-office working is the default
How to apply
This role is advertised on a confidential basis. Applications are made through theHRkey: select “Apply through theHRkey”, add your CV and answer a short set of questions about your experience. Your application is received straight away and we will follow up on next steps.