Machine Learning Engineer, AI Safety

About the role

The work covers content safety, product security, fairness and robustness for language and multimodal models, with metrics and tooling that other ML teams rely on. Datasets, evaluations and MLOps practices are central to it.

At a glance

What you will do

• Build datasets and models for training and evaluating content safety, product security, robustness and fairness systems
• Research and apply methods for detecting and reducing bias in LLMs and in systems built on them such as RAGs
• Define and track metrics for responsible model behaviour and usage
• Follow MLOps practice for automation, monitoring, scale and safety, and build safety tools for other ML teams
• Work with engineers, data scientists and researchers on content safety and fairness problems

What they look for

• A master's or PhD in computer science, electrical engineering or a related field, or equivalent experience
• At least 2 years developing and deploying machine learning models in production
• Solid grounding in ML principles and algorithms, with hands-on Python and frameworks such as Keras or PyTorch
• Practice with large multimodal datasets and models, plus strong analytical and communication skills

Nice to have

• Experience aligning or fine-tuning LLMs, including vision-language or any-to-text models
• Knowledge of generative AI security risks such as prompt instability, model extraction and adversarial robustness