LLM Machine Learning Engineer, Models and Agent Science

What you'll do

• Develop and optimize large language models (LLMs) for both on-device and server-based applications.
• Evaluate and reproduce recent research to determine its practical viability for Apple products.
• Implement post-training techniques to improve model quality, performance, and behavior steering.
• Build agent harnesses and develop interpretability or stress-testing algorithms.
• Translate cutting-edge machine learning research into scalable features within the product context.
• Conduct scientific investigations and experiments while managing real-world constraints and uncertainty.
• Design rigorous experiments and ablation studies to benchmark LLM optimizations.
• Utilize Python, UNIX, and agentic coding tools to develop and deploy machine learning capabilities.

What we're looking for

• A BS and 5+ years of experience, MS and 3+ years of experience, or PhD and 1+ year of experience is required.
• Experience in post-training large language models (LLMs), including quality/performance optimization, interpretability, and steering model behavior.
• Strong software engineering background with proficiency in Python and UNIX environments.
• Demonstrated ability to use agentic coding tools and conduct research on neural network optimization or model training.
• Proven ability to evaluate recent research and translate it into viable products under real-world constraints.
• Experience working with large-scale compute infrastructure and experimental rigor in benchmarking LLM optimizations.
• A publication record at top AI/ML venues is preferred.
• Ability to drive scientific investigations and experiments while managing uncertainty in a research environment.