Описание
Apple builds AI systems that shape experiences for billions of people, with a commitment to privacy, performance, and craft. Its AI & Data Platforms team develops generative AI systems and products at global scale.
Задачи
•Lead the end-to-end development and productionisation of LLM-based systems, from upstream training and reinforcement learning through fine-tuning, alignment, and deployment of globally scaled products
•Design and implement LLM evaluation and benchmarking frameworks to assess model quality, safety, bias, latency, and cost-efficiency
•Architect production inference infrastructure, including model optimisation, quantisation, and efficient serving strategies
•Drive model customisation and adaptation strategies, including prompt engineering, retrieval-augmented generation, and parameter-efficient and full fine-tuning
•Build end-to-end AI-powered products and features, owning delivery from problem definition and prototyping through production release across Swift, Java, and Python codebases
•Establish engineering standards across the ML development lifecycle, including testing, reproducibility, monitoring, documentation, and CI/CD for model and data pipelines
•Partner with research, product, design, and platform teams to turn emerging capabilities into scalable, user-centric solutions
•Mentor ML engineers, raise technical quality, and foster rigorous experimentation and engineering craft
Требования
•Extensive hands-on Machine Learning engineering experience and a track record of shipping ML-powered products at scale
•Practical expertise in LLM fine-tuning, alignment, and customisation, including RLHF, LoRA, QLoRA, prompt optimisation, and LLM evaluation and benchmarking
•Strong software engineering proficiency in Python, Swift, and Java
•Experience building and operating enterprise-grade ML pipelines in cloud or on-prem environments
Будет плюсом
End-to-end AI product delivery, published papers in top ML/Statistics/Maths/computer science conferences, LLM pre-training, reinforcement learning for model alignment, safety and red-teaming, agentic frameworks, multimodal AI systems, standalone AI-native products, open-source contributions, research or patents, inference optimisation, data engineering, architectural direction, cross-team alignment, and mentoring senior engineers.
