Machine Learning Engineer, Proactive

Apple; Cupertino (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Apple seeks a Machine Learning Engineer to design, train, and deploy large language models and on-device AI systems that power intelligent search and personalized experiences across Apple’s ecosystem.

You will run end-to-end projects from research to production, optimize models for latency and privacy, and collaborate with engineers, researchers, product managers, and designers to bring new AI capabilities to users.

Qualifications

• Bachelor's degree in CS/ML or related field.
• Experience with transformers, LLMs, or large language models.
• Proficiency in PyTorch, JAX, or TensorFlow.

Responsibilities

• Design, train, and deploy transformer-based models for on-device use.
• Build semantic retrieval, embedding, reranking, and retrieval-augmented systems.
• Collaborate with engineers, researchers, PMs, and designers to ship features.
• Evaluate models with offline and online metrics, optimize latency and privacy.

Skills

Machine learning & NLP

Transformers & LLMs

Python / C/C++

On-device AI

Evaluation & research

Education

Bachelor's degree in CS/ML

Advanced degree preferred

Job description

Summary

At Apple, machine learning powers experiences that anticipate what people need before they ask. We're looking for a Machine Learning Engineer to help build the next generation of intelligent search and AI experiences technology that understands user intent, context, and personal information while preserving privacy. In this role, you'll design, train, fine-tune, optimize, and deploy large language models, semantic retrieval systems, and ranking models that power relevant, personalized, and context-aware experiences across Apple's ecosystem.

Description

You'll design, train, fine-tune, and optimize transformer-based language models and foundation models for efficient on-device deployment, and build semantic retrieval, embedding, reranking, and retrieval-augmented generation systems that improve search quality and AI-powered experiences. You'll develop models for query understanding, intent prediction, personalization, retrieval, and ranking, while researching new approaches to LLM fine-tuning, knowledge distillation, model compression, quantization, and low-latency inference. You'll explore techniques for adapting large foundation models into smaller, highly capable models that can operate efficiently under on-device memory, compute, power, and latency constraints.You'll partner with engineers, researchers, product managers, and designers to bring new AI capabilities from research into production, driving technical strategy and leading projects from early exploration through large-scale deployment. This is an opportunity to explore new applications of foundation models, multimodal AI, agentic retrieval, and personalized intelligence, shaping the next generation of proactive and intelligent user experiences.

Key Responsibilities

Build semantic retrieval, embedding, reranking, and retrieval-augmented generation systems, along with models for query understanding, intent prediction, personalization, retrieval, and ranking.Analyze search relevance and user behavior to design evaluation methodologies, offline benchmarks, and online metrics that measure retrieval quality, ranking, personalization, and language model performance.Build scalable experimentation and evaluation pipelines for LLMs and search models, including model quality, robustness, latency, efficiency, and end-to-end product metrics.Design, train, fine-tune, distill, and optimize transformer-based language models and foundation models for efficient on-device deployment.Develop LLM fine-tuning and post-training approaches, including supervised fine-tuning, instruction tuning, preference optimization, parameter-efficient fine-tuning, and task-specific adaptation.Research and prototype approaches for on-device generative AI, including knowledge distillation, model compression, quantization, pruning, and low-latency inference.Develop techniques to transfer capabilities from large foundation models into compact on-device models while balancing model quality, latency, memory footprint, power consumption, and compute constraints.Partner with engineers, researchers, product managers, and designers to bring AI capabilities from research into production, driving technical strategy across projects and exploring new applications of foundation models, multimodal AI, agentic retrieval, and personalized intelligence.

Minimum Qualifications

• Bachelor degree in Computer Science, Machine Learning,Artificial Intelligence, or a related field.
• Background in machine learning, deep learning, naturallanguage processing, information retrieval, search,recommender systems, or generative AI.
• Experience training, fine-tuning, or deploying transformer-based models and large language models.
• Experience with modern deep learning architectures andtechniques, including transformers, embeddings,representation learning, and neural ranking.
• Programming skills in Python and/or C/C++, with experiencebuilding production-quality software using modern machinelearning frameworks such as PyTorch, JAX, or TensorFlow.
• Ability to work onsite in Cupertino, California, in accordancewith Apple's applicable work policies.

Preferred Qualifications

• Master's or Ph.D. in Computer Science, Machine Learning, Artificial Intelligence, or a related field.
• Experience optimizing machine learning models for resource-constrained environments, including knowledge distillation, model compression, quantization, and pruning.
• Experience with on-device machine learning or edge AI, or mobile inference frameworks, including optimizing models for latency, memory, compute, and power constraints.
• Experience distilling capabilities from large foundation models into small language models or task-specific models for efficient inference.
• Experience building retrieval-augmented generation, vector search, embedding retrieval, neural reranking, or semantic search systems.
• Experience with query understanding, query rewriting, intent classification, personalized retrieval, learning-to-rank, or recommendation models.
• Experience working with transformer architectures and foundation model families such as BERT, T5, Llama, Gemma, Mistral, or related architectures.
• Experience evaluating language models, designing AI quality metrics, and building automated and human-in-the-loop evaluation pipelines.
• Experience building large-scale production search, recommendation, personalization, or generative AI systems.
• Familiarity with multimodal foundation models, tool use, agentic AI, or agentic retrieval systems.
• Strong understanding of the tradeoffs among model quality, latency, memory, power consumption, privacy, and reliability for production on-device AI systems.
• Ability to prototype new ideas, conduct rigorous experiments, solve ambiguous technical problems, and translate research advances into production-quality machine learning solutions.