AIML - ML Researcher, AFM

Apple • Cupertino, CA

1d

About The Position

The AIML ML Researcher will work on foundation models across the training lifecycle, including pre-training and mid-training approaches that connect general capabilities with task-specific performance. The team engineers models for Apple silicon and optimizes them for private, personal experiences integrated into the operating system. Research areas include reward modeling that resists reward hacking, sparse and delayed rewards in agentic settings, and reliable model alignment for both open-ended creative tasks and precise action-taking workflows.

The researcher will collaborate with a close-knit, fast-growing team of engineers and scientists on challenging problems in foundation models and deep learning, including end-to-end voice interactions, natural language processing, and combining learning with knowledge. The work emphasizes applying deep learning research to real-world product use cases.

Requirements

• Demonstrated expertise in deep learning, supported by publications at relevant conferences or a track record of applying deep-learning techniques to products.
• Proficient programming skills in Python and one deep-learning toolkit such as JAX, PyTorch, or TensorFlow.
• Ability to work in a collaborative environment.
• PhD in Computer Science or a related technical field, or equivalent practical experience.

Nice To Haves

• Experience building end-to-end voice models.
• Experience in speech recognition and generation.
• Experience in reinforcement learning and on-policy distillation.
• Experience in post-training and mid-training large language models.
• Experience working with large language models.

Responsibilities

• Conduct research and development across the foundation-model training lifecycle, including pre-training and mid-training.
• Develop approaches that bridge general model capability and task-specific performance.
• Work on reward modeling, including methods to resist reward hacking.
• Address sparse and delayed rewards in agentic settings.
• Improve reliable model alignment for creative and action-taking workflows.
• Collaborate with engineers and scientists on foundation-model and deep-learning problems, including voice interactions, natural language processing, and combining learning with knowledge.

What This Job Offers

• Job Type: Full-time
• Education Level: Ph.D. or professional degree