Siri AI talks with hundreds of millions of people in dozens of languages, and each market expects it to understand its language, culture, and context as well as a local would. It takes the right data, careful evaluation, models built for each locale, and systems that can do all of this across many markets at once.
The Global Siri team builds the next generation of Siri's AI agents for users around the world. We're looking for a strong AI engineer who works across the whole ML lifecycle (data, modeling, evaluation, and deployment) and uses agentic tools every day to get more productive to make Siri great in every language.
As an applied AI Engineer on Global Siri, you'll work closely with Siri's modeling, evaluation, data science, and deployment teams. You'll also partner with Apple's Modeling, ML platform, and safety teams. You'll design, build, and ship the data, models, and systems that bring Siri's AI features to new languages and markets and keep raising quality in the markets we already serve.
Multilingual data as a product: Build scalable process, agent and pipeline that could curate, filter, deduplicate, augment, and version multilingual data for Siri AI. This includes fine-tuning sets, tool-use trajectories, evaluation sets, and synthetic data.
Modeling for global markets: Use GenAI, NLP, and ML methods to improve how Siri understands and responds in every language, and find and fix quality gaps that differ by locale.
Evaluation at scale: Build systems that run large multilingual evaluation suites for agentic models, so every model candidate is measured quickly and consistently in every market.
Agentic automation: Use agents and AI-assisted coding to automate internationalization work, such as data generation, annotation, error analysis, and localization QA.
Data-driven prioritization: Use metrics and analysis to decide which markets, languages, and failure types to tackle first.
Communicating results: Share your findings, tradeoffs, and recommendations clearly with executives and partner teams to shape priorities and launch decisions.
B.S. or M.S. in Computer Science or a related field, or equivalent experience
Strong software engineering fundamentals and proficiency in Python
Hands-on experience across the ML lifecycle (data, training, evaluation, deployment).
Strong understanding of LLM, NLP, ML and their application.
Hands-on experience with ASR, acoustic triggering, or TTS in international locales, including accents, code-switching, low-resource languages, and locale-specific metrics
Extensive hands-on experience with agentic AI-assisted coding and context engineering
An analytical, data-driven mindset and experience.
Experience building and debugging production services or data products end to end
Excellent communication skills and the ability to work across teams in an English-speaking environment
Strong written and verbal communication skills
Agentic coding and GenAI as your daily way of working, including building your own agents or workflows that automate real engineering and data tasks
Experience building synthetic data generation pipelines and using them to launch features or improve product metrics, especially where real data is scarce
Experience with large-scale data or ML infrastructure (e.g. Spark, Kubernetes) and LLM evaluation, such as LLM-as-judge
A track record of shipping user-facing ML features in multiple languages or markets
Job description provided by Apple.
View the original posting
