Red hat Senior Machine Learning Engineer Responsible AI - Waterford 

Ireland 

611194199

Job Responsibilities

• Architect and implement comprehensive safety systems for LLM deployments, including content filtering, output validation, and alignment techniques

• Design and develop robust guardrailing frameworks that enforce model behavioral boundaries while maintaining performance and user experience

• Lead the development of monitoring systems to detect and mitigate potential model hallucinations, harmful outputs, and alignment drift in production

• Build and maintain evaluation frameworks for assessing model safety, including automated testing pipelines for toxicity, bias, and harmful behavior

• Develop prompt engineering systems and safety layers that ensure reliable and controlled LLM outputs across different use cases and deployment scenarios

• Implement fine-tuning and human preference alignment pipelines with a focus on maintaining model alignment and improving safety characteristics

• Design and deploy systems for LLM output validation, including fact-checking mechanisms and source attribution capabilities

• Lead technical initiatives around model interpretability and transparency, including debugging tools for understanding model decisions

• Collaborate with policy and safety teams to translate safety requirements into technical implementations and measurable metrics

Requirements

• 5+ years of ML engineering experience , with 3+ years specifically working with transformer-based models and LLMs

• Deep expertise in prompt engineering , instruction tuning , or human preference alignment techniques

• Strong background in implementing AI safety mechanisms and guardrails for production LLM systems

• Experience with LLM evaluation frameworks and safety metrics

• Proven track record of building production-grade systems for model monitoring and safety enforcement

• Strong programming skills in Python and experience with modern LLM frameworks (PyTorch, Transformers, etc.)

• Experience implementing content filtering and output validation systems

• Understanding of AI alignment principles and practical safety techniques

The following will be considered a plus:

• Experience with constitutional AI and alignment techniques

• Background in implementing human preference alignment pipelines and fine-tuning large language models

• Familiarity with LLM deployment platforms and serving infrastructures

• Experience with model interpretation techniques and debugging tools for LLMs

• Knowledge of AI safety research and current best practices

• Understanding of adversarial attacks and defense mechanisms for language models

• Experience with prompt injection prevention and input sanitization techniques

• Background in implementing automated testing systems for model safety

• Advanced degree in Computer Science, ML, or related field with focus on AI safety