Senior Machine Learning Infrastructure Engineer, Music
August 15
🔄 Hybrid – London
Description
• Build massive data processing pipelines to run ML models on audio at petabyte-scale • Work cross-functionally with other engineers, research scientists, and product managers to build audio ML research into Spotify’s products to serve hundreds of millions of users • Build best-in-class infrastructure and tooling to support and accelerate model training • Work closely with research scientists to debug, optimize, and convert ML models for deployment • Debug problems across the entire stack, from low-level debugging to high-level system design • Collaborate with customer teams to deploy our research in products around Spotify, in backend, data, mobile, core, and other code bases • Scope the feasibility of projects through quick prototyping to assess performance, quality, time and cost • Help maintain crucial pieces of Spotify-owned open-source audio software infrastructure like Pedalboard, Basic Pitch, and more
Requirements
• You have professional experience working with machine learning systems • You have a very strong grasp of Python, but are happy to work in languages like Java, Scala, C++, and others as necessary • You have very strong communication skills and can collaborate effectively with ML researchers, product stakeholders, and executives across various geographies • You have strong systems fundamentals and can debug problems down to the operating system level • You have a good understanding of performance and can design and engineer systems that scale without breaking the bank • You have experience debugging, profiling, optimizing, or deploying ML models • You have worked with cloud platforms like GCP, AWS, or Azure • You have proven experience implementing and maintaining large-scale production software systems • You are interested in learning more about audio processing and music information retrieval and you're excited about building products that use such technologies • You are willing to go on call to support the systems we build, and are willing to build systems for reliability to avoid on-call fatigue. (The team averages 1 incident per quarter.) • Experience with deep learning techniques for content based processing (audio, image, video data) is a plus
Apply Now
