AIML - Agentic LLM Intern - Siri

Summary

Posted: Oct 24, 2024
Weekly Hours: 40
Role Number:200575609

The AI/ML Global Product Experience team creates groundbreaking user experiences across Siri markets using natural language processing (NLP), machine learning (ML) and modern software development techniques. The features we build are redefining how hundreds of millions of people across all Siri languages are connected to the information they are looking for and the apps they love to use through various devices. As part of this group, you will work in one of the most exciting environments with state-of-the-art ML and NLP models applied to production problems. We build ML technologies that scale to all Siri languages. You will have the opportunity to innovate with multilingual NLP models to help Siri understand different languages and user queries better.

Description

In Global Siri, we are responsible for the end-to-end user experience in all Siri markets. This means we build new features and scale them across Siri languages. As an Intern in Global Siri, you will help us build ML solutions to scale new features, utterance understanding and natural language generation faster across different languages. You should be able to work with large scale systems, write high quality code and are comfortable in contributing to existing systems and developing new ones. Communication skills will be required to coordinate work across multiple teams. Specific responsibilities include: * Work on cutting-edge research at the intersection of Agents & Multilinguality * Understand how agentic LLMs learn capabilities across different languages * Devise automated approaches to evaluate the performance of agents across multiple languages * Find innovative solutions to scale fine-tuning approaches to multiple languages * Communicate and share your findings with the larger ML community and Global Siri teams at Apple

Minimum Qualifications

• Expertise in state-of-the-art LLMs and in-depth knowledge of the latest advancements in the field
• Deep understanding of fine-tuning and reinforcement learning techniques
• Experience with or strong interest in agents, tool-use, planning and multi-step reasoning
• Experience with multilingual data and understanding of the complexities and tradeoffs involved when scaling to non-English languages

Preferred Qualifications

• Currently working towards a PhD degree in Computer Science or related technical field
• Proficiency in Python and modern ML frameworks such as TensorFlow, PyTorch or JAX
• Publication record in relevant conferences demonstrating ability to conduct innovative research in deep learning or a track record in applying deep learning techniques to products