We build and develop the core language and machine translation models that power Localization across Services in an efficient and scalable manner — and the production systems that put those models in front of users. We work on a wide spectrum of approaches, including agentic workflows, foundation modeling, deep learning, model compression, and transfer learning. We also build the systems that power Apple Music lyrics translations and lyrics transliterations (phonetic pronunciation). This position spans applied modeling and the software engineering needed to ship at scale, offering a unique chance to work where Localization meets state-of-the-art, large-scale software development.

Does this sound like you? Join our team!

Minimum Qualifications
BS/MS/PhD in a quantitative field (Computer Science, Math, Statistics, Physics, etc.) and 5+ years of relevant experience.
Strong software engineering fundamentals and proficient programming skills in Python, with experience writing and maintaining production-quality code.
Hands-on experience with deep learning toolkits such as JAX, TensorFlow, or PyTorch.
Proven track record training or deploying large models in production.
Experience building or operating large-scale, distributed production systems.
Deep understanding of Deep Learning, Large Language Models (LLMs), and Natural Language Processing (NLP).

Preferred Qualifications
PhD in a quantitative field, or equivalent depth in machine translation, multilingual NLP, or applied LLM research.
Experience with machine translation and translation-quality evaluation (e.g., COMET, BLEU, human evaluation).
Experience optimizing model serving and inference — quantization, distillation, or compression — for low-latency, high-throughput deployment.
Familiarity with MLOps and model-deployment infrastructure.
Experience building LLM-based agents, including the ReAct pattern in agentic workflows.