Senior Machine Learning Engineer

Apple•Seattle, WA

47d•$171,600 - $258,100•Onsite

About The Position

Research and develop advanced machine learning solutions for natural language processing (NLP), audio, computer vision, and multi-modal applications. Work closely with crossfunctional partners, including product teams, to productionize large-scale ML solutions. Provide technical guidance to improve workflows for training, evaluation, model optimization, and deployment. Contribute modular and performant Python code to internal PyTorch model development frameworks; develop, release, and support new features for these projects. Implement and maintain state-of-the-art neural networks architectures in internal frameworks. Architectures include large language models (LLMs), audio models, vision transformers, and multimodal generative AI models. Design, implement, and maintain production workflows for large-scale distributed training using advanced techniques such as fully sharded data parallelism and model parallelism. Implement and support tooling for evaluating the quality of large deep learning models. Research and evaluate the latest techniques for machine learning model development; incorporate them into production workflows. Contribute to the organization’s technical roadmap by identifying areas for innovation in machine learning solutions. Provide technical leadership to improve workflows for training, evaluation, model optimization, and deployment. Mentor and guide junior engineers and interns in best practices for machine learning model development.

Requirements

Master’s degree or foreign equivalent in Computer Science, Engineering, or related field and 3 years of experience in the job offered or related occupation.
2 years of experience with applying ML fundamentals to make quantifiable improvements to prior solutions in the domains of ML systems, natural language processing, computer vision, and/or audio processing.
2 years of experience utilizing modern ML frameworks, including PyTorch or JAX, to train and evaluate generative AI models for natural language processing, computer vision, or audio applications.
2 years of experience applying software engineering fundamentals, with a strong focus on data structures and algorithms, to develop efficient and maintainable Python code for ML applications.
2 years of experience using collaboration tools including git and GitHub to contribute to a large codebase with 3+ collaborators.
Experience and/or education must include implementing and maintaining tooling for industrial ML model development using Python and PyTorch or JAX.
Experience and/or education must include implementing state-of-the-art ML model architectures in PyTorch or JAX, including transformer models for text, vision, or audio applications.
Experience and/or education must include applying fundamentals of ML systems to scale up ML training to large models (at least 1 billion parameters) using fully sharded data parallelism (FSDP) or model parallelism.
Experience and/or education must include applying parameter efficient fine-tuning techniques, including LoRA (low-rank adaptation), to fine tune foundation models in a computationally efficient way.
Experience and/or education must include using distributed machine learning frameworks, including torch.distributed, DeepSpeed, or Ray, to build machine learning infrastructure or platforms.
Experience and/or education must include evaluating the quality of generative AI models using evaluation frameworks including lm-evaluation-harness.

Responsibilities

Research and develop advanced machine learning solutions for natural language processing (NLP), audio, computer vision, and multi-modal applications.
Work closely with crossfunctional partners, including product teams, to productionize large-scale ML solutions.
Provide technical guidance to improve workflows for training, evaluation, model optimization, and deployment.
Contribute modular and performant Python code to internal PyTorch model development frameworks; develop, release, and support new features for these projects.
Implement and maintain state-of-the-art neural networks architectures in internal frameworks.
Design, implement, and maintain production workflows for large-scale distributed training using advanced techniques such as fully sharded data parallelism and model parallelism.
Implement and support tooling for evaluating the quality of large deep learning models.
Research and evaluate the latest techniques for machine learning model development; incorporate them into production workflows.
Contribute to the organization’s technical roadmap by identifying areas for innovation in machine learning solutions.
Provide technical leadership to improve workflows for training, evaluation, model optimization, and deployment.
Mentor and guide junior engineers and interns in best practices for machine learning model development.