Note: The job is a remote job and is open to candidates in USA. Harnham is an early stage AI startup focused on improving the safety and reliability of AI systems in production environments. They are seeking an ML Engineer to take full ownership of building a critical machine learning system, designing and delivering core training and remediation loops.

Responsibilities

• Build and own end to end training and remediation pipelines for AI systems, from signal detection through to deployment and monitoring
• Develop reinforcement learning and preference optimisation workflows, including approaches such as DPO and RLHF
• Translate behavioural issues into structured training data and deploy effective model improvements
• Create scalable evaluation and testing systems to ensure fixes perform reliably in production
• Work across multiple models and deployments, turning one off fixes into reusable tooling
• Improve system performance, reliability, and efficiency as the product evolves

Skills

• Strong commercial experience in machine learning engineering, with deep expertise in Python and PyTorch
• Proven ability to train, fine tune, and deploy models in production environments
• Hands on experience with reinforcement learning or preference optimisation techniques such as DPO, RLHF, or PPO
• Strong understanding of end to end ML pipelines, from data preparation through to monitoring
• High standards for code quality, architecture, and readability
• Comfortable operating in an early stage environment with ambiguity and ownership
• Experience designing evaluation systems or working on AI safety, trust, or security is beneficial

Benefits

• Meaningful early stage equity package
• Comprehensive benefits including healthcare, retirement, and generous holiday allowance
• The opportunity to build a core platform from scratch and shape technical direction
• Close collaboration with founders and engineering leadership

Company Overview

Company H1B Sponsorship