LLM Machine Learning Research Engineer

Apple is seeking a Research Engineer to join our Foundation Model Preparation and Algorithm Team. We are looking for all levels of talent to bring innovative AI research into Apple products.

We are looking for strong ML applied scientists and engineers to build groundbreaking AI infrastructure and algorithms. This infrastructure will power the optimization of Apple Foundation models, including on-device and server Apple Intelligence models. My team is directly responsible for general model capability, use-case-oriented post-training, and also the feature delivery for Apple Intelligence.\n\nYou should be a strong scientist and/or engineer who has a background in building state-of-the-art LLMs. Your work will have a direct impact on billions of Apple clients. You will collaborate with world-class talent in LLM training, on-device and server optimization, ML tools/platforms, datasets, and evaluation. You will develop reliable and scalable pipelines and algorithms, such as:Model optimization pipelines, State-of-the-art optimization algorithms, State-of-the-art post-training techniques.

Experience developing, optimizing, or training large language models (LLMs), large foundation models, or generative AI models.\nSoftware engineering skills in Python and general-purpose system administration and infrastructure management abilities.\nHistory of applied research in the neural network model life cycle, training, or a related application area.\nExperience with languages like Python, C/C++.\nTrack record of driving scientific investigations and experiments, and overcoming obstacles and uncertainty in a research environment.\nBS degree and 3+ years of proven experience.

Publication record at top AI/ML venues.\nExperience with LLM LoRA fine-tuning, neural network optimization (e.g., quantization, palettization).\nExperience with LLM pre-training or post-training.\nExperience with on-device/server scale deployment.\nInfrastructure management and debugging experience.\nExperimental rigor when training/evaluating LLMs for the purpose of benchmarking LLM optimization algorithms.\nStrong communication and accountability skills; a hard-working, strong work ethic, and collaboration abilities.\nPh.D. in a related field.