Summary

The ML systems engineer will build training, data, and evaluation infrastructure for Cosine’s software-engineering models. This is a fully in-office role in London.

Responsibilities

• Build supervised fine-tuning and reinforcement-learning pipelines
• Generate, filter, and sample training datasets
• Deploy containerised services for RL infrastructure
• Create repository environments for agents to test code
• Extend evaluation suites for software-engineering models
• Analyse model failures and improve reward functions