Machine Learning Engineer – Lumen Enterprise Models (SWE-focused LLMs)
Salary not listed • London Office • Full-time
Posted 16d ago
Job Location
London Office
Tech Stack
Remote Work Policy
On-site
Employment Type
Full-time
Categories
LLM Engineer
About the job
Cosine is developing autonomous software-engineering systems for organizations that prioritize reliability, security, and control. Our Lumen Enterprise models are designed to understand and modify real codebases, utilizing tools, running tests, and producing maintainable software. We are seeking a Machine Learning Engineer to enhance and train these models across the entire training stack, including pre-training, supervised fine-tuning, and reinforcement learning. This role offers exposure to all facets of the model training lifecycle, from performance profiling to implementing new RL algorithms, and requires a hands-on approach with a focus on software quality.
Responsibilities
- Work across the full training lifecycle, running pre-training, supervised fine-tuning, and reinforcement-learning experiments.
- Optimize training methods at scale on frontier models.
- Diagnose and fix performance bottlenecks and communication overhead in distributed training environments.
- Train models on long-context, multi-file, and repository-level tasks.
- Develop RL and RLVR loops using signals from execution, tests, linters, static analysis, and other engineering feedback.
- Investigate model failure modes such as poor planning, brittle fixes, reward hacking, and regressions.
- Collaborate with infrastructure, platform, and data teams.
Requirements
- Strong experience training deep-learning models in a production or research environment.
- Deep practical experience with PyTorch or a comparable framework.
- Experience with distributed training and multi-GPU systems (FSDP, Deepspeed).
- Strong software-engineering fundamentals and a high bar for correctness and maintainability.
- Clear communication and the ability to explain technical trade-offs.
- Meaningful experience with at least one of: pre-training, SFT, preference optimisation, RLHF, or RLVR.
Benefits
- Salary and Equity options available for discussion