ML Systems Engineer for SWE-Focused LLMs (On-site & Equity)

Cosine; Greater London; On-site; GBP 80,000 - 110,000; Full time

Benefits offered by this job

Competitive salary

Equity / share options

30 days’ holiday + bank holidays

Dog-friendly office

Weekly team breakfast & lunch

Monthly socials

Pension

High-quality equipment

Job summary

Cosine is building autonomous AI engineers that plan, write and ship code inside real development workflows. You will build environments in which models learn to write software, develop pipelines for training data, and run fine-tuning and RL workloads that shape model behavior.

This is a hands-on role at the intersection of ML and software engineering, requiring production-grade tooling, PyTorch, Docker/Kubernetes, and data-centric optimization.

Qualifications

• Typically 3–5 years of software engineering or ML infra experience.
• Proficiency in Python and Go; production-grade code.
• Experience with Docker and Kubernetes.
• Understanding data pipelines and model training.

Responsibilities

• Build end-to-end training systems for SWE models in Lumen.
• Create data-generation and curation pipelines for training datasets.
• Develop RL infrastructure and containerised services for training and evaluation.
• Improve evaluation suites for code models and measurement of model quality.
• Shape training directions with research, infra and product teams.

Skills

Python

Go

Software engineering

Code quality

Ownership

Debugging

Tools

Docker

Kubernetes

PyTorch

TensorFlow

JAX

Job description

Cosine is building autonomous AI engineers that plan, write and ship code inside real development workflows. You will build environments in which models learn to write software, develop pipelines for training data, and run fine-tuning and RL workloads that shape model behavior.

This is a hands-on role at the intersection of ML and software engineering, requiring production-grade tooling, PyTorch, Docker/Kubernetes, and data-centric optimization.