The Software Localization team at Apple needs to develop the next generation of ML tooling and end-user features for international locales to ensure Apple and third-party developers using Xcode can deliver world-class software products to millions of users around the world.
Requirements
- Experience with Machine Learning (ML), with a particular emphasis on Large Language Models (LLMs) and Natural Language Processing (NLP)
- Proven ability to comprehend, interpret, and apply cutting-edge research into consumer-oriented products
- Comprehensive knowledge and hands-on experience with fine-tuning approaches and training models
- Experience adapting pre-trained LLMs for downstream tasks
- Proficiency using open-source ML toolkits and frameworks (e.g. PyTorch, TensorFlow)
- Strong programming skills in Python, Swift, C or C++
- Proficiency with Apple’s development APIs (SwiftUI, Foundation, etc.)
Responsibilities
- Leading the exploration and application of Large Language Models (LLMs) for performing translation tasks and Apple product features
- Translating the latest research into high-quality systems and models that can be practically applied to further improve machine translation or Apple product features
- Actively engaging in all aspects of model development, from ideation, training, experimentation to deployment.
- Developing and maintaining features, frameworks and tools to help facilitate the software translation process
- design, implement and qualify localization features, processes and tooling across a wide variety of Apple’s products.
- architect the future of localizing models and machine translation.
Other
- Collaborating with translation, localization quality, project management and engineering teams to develop and implement solutions
- Experience in software localization / internationalization
- Familiarity with localization standards and technologies (Unicode, XLIFF, ICU, etc.)
- Native-level foreign language skills in reading, writing and speaking
- Comfortable working cross-functionally and cross-culturally
