AIML - ML Engineer, Responsible AI and Safety
Software Engineering, Data Science
USD 184,700-324,800 / year + Equity
Posted on Nov 20, 2026
Join our team at the forefront of defending Apple's ecosystem. Our Mitigation Engineering team builds the large-scale machine learning systems that protect millions of users from emerging threats and ensures the integrity of our products. We are looking for an experienced, senior Applied ML Engineer who has a proven track record of shipping production models. The ideal candidate is passionate about tackling complex safety and security challenges using state-of-the-art techniques, including methods that draw on a model's internal representations for mitigation purposes. In this role, you will design, build, and deploy the critical machine learning systems that are foundational to the safety of Apple's products, all while upholding our deep commitment to user privacy.
As a senior engineer on this team, you will own the full lifecycle of our abuse detection and safety guardrail models. You will partner closely with software, platform, and product teams to deploy robust, scalable defenses. We believe the most effective security systems are built by engineers who can translate adversarial insights into production-ready code. Your work will directly contribute to the architecture of Apple's AI platform and protect users from real-world harm.
- Design, build, and deploy production-grade ML models to detect and mitigate risks across multiple modalities (text, image, audio) and languages
- Own the modeling end to end: from prototyping and data analysis to deployment, monitoring, and the continuous improvement of models in production - this includes decisions such as probe based mitigation architectures, layer selection, calibration, and robustness
- Drive the data strategy to continuously improve model performance by analyzing distribution gaps, contributing to synthetic data pipelines, and creating automated annotation systems
- Work with platform and inference teams on how guardrail models are hosted across different model architectures and deployment targets, and shape modeling choices around latency, memory, and portability
- Collaborate with cross-functional partners in engineering, research and product teams to define project requirements, establish technical direction, and deliver robust safety solutions
- Proven experience shipping machine learning models to production. You have owned the end-to-end lifecycle of a model, from development to deployment and maintenance
- Strong familiarity with research fundamentals, machine learning principles, and development methodologies around LLMs, foundation models, including how transformer internal states (activations) can be used to detect safety-relevant behavior
- Proficient programming skills in Python and deep learning toolkits (e.g. JAX, PyTorch, Tensorflow)
- Ability to work with sensitive and offensive content as part of building robust security and abuse detection systems
- BS, MS or PhD in Computer Science, Machine Learning, or related fields or an equivalent qualification acquired through other avenues
- 5 years of experience shipping machine learning models to production systems
- Hands-on experience with activation probing, representation-level or interpretability methods, or fine-tuning or aligning large language models for security or safety applications
- Interest in or experience with deploying models and lightweight model heads across different architectures and hardware, such as on-device and server-side
- Experience driving technical projects and collaborating with large, diverse, cross-functional teams