SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: GBP 80,000 - 110,000 / annual
Cosine is building autonomous AI engineers that plan, write and ship code inside real development workflows. The company has achieved a 72% score on OpenAI's SWE-Lancer benchmark, placing it among the strongest real-world software-engineering AI systems evaluated.
You will join as an ML Systems Engineer working at the intersection of machine learning, software engineering, data and infrastructure. This is a highly hands-on role focused on building the environments in which Lumen models learn to write software, developing pipelines that generate and curate training data, and running fine-tuning and reinforcement-learning workloads that shape model behaviour.
Key responsibilities include:
- Contributing to end-to-end training of software-engineering models, implementing supervised fine-tuning pipelines using curated code and conversation datasets.
- Building reinforcement-learning loops in which models write code, run tests and use development tools.
- Developing custom PyTorch dataloaders, training objectives and evaluation workflows.
- Creating synthetic data-generation pipelines and designing systems for producing, filtering, transforming and sampling large-scale datasets.
- Designing, building and deploying containerised services that support model training and evaluation using Docker and Kubernetes.
- Building environments where agents can modify repositories, run tests, use tools and receive meaningful feedback.
- Maintaining and extending evaluation suites for code models, including unit tests, benchmark suites and repository-level tasks.
- Analysing model failure modes and developing better ways to measure code quality, maintainability and correctness.
- Working with research, infrastructure and product teams to identify high-value training problems and turn broad goals into clear experiments.
- Developing opinionated reward functions for software-engineering agents.
You will work closely with ML researchers, infrastructure engineers and product teams to decide what Lumen should learn next, how to generate the right training data, how to design environments that reflect real software-engineering work, and how to reward models for producing useful, maintainable code.
The role is based in Cosine's Hoxton office in London with full in-office working as default, five days a week. The company emphasizes close collaboration, fast feedback and shared context.
QUALIFICATIONS & REQUIREMENTS:
You should have:
- Typically 3–5 years of experience in software engineering, ML infrastructure, data engineering, applied machine learning or a closely related field, or equivalent evidence of strong technical ability.
- Strong software engineering fundamentals: ability to read, debug and write non-trivial production code; working primarily in Python and Go; caring about correctness, maintainability and code quality as much as model metrics; taking ownership of ambiguous technical problems.
- Practical experience with at least one of PyTorch, TensorFlow or JAX; implementing custom training loops, losses or dataloaders; understanding the relationship between data, objectives, evaluation and model behaviour.
- Experience with Docker and container-management or orchestration platforms such as Kubernetes.
- Experience with at least one major cloud platform (GCP, AWS or Azure).
- Understanding of how to build services that are reliable, observable and straightforward to operate.
- Experience working with large-scale datasets and object storage; understanding of sharding, filtering, sampling and dataset versioning.
- The ability to explain technical decisions and experimental results clearly; comfort documenting trade-offs and walking others through your reasoning.
Nice to have: synthetic data generation for code or language models, training LLMs in distributed environments, reinforcement learning or preference optimisation, data tooling such as SQL or DuckDB, LLM-as-a-judge systems, reward-hacking detection, open-source contributions to LLM tooling or ML infrastructure, experience with repository-level code tasks or software-engineering benchmarks.
