Machine Learning Data Researcher

Apple's Hardware group is home to some of the world's most advanced technologies - powering the experiences that define Apple products. We are looking for a talented and curious ML Data Researcher to join our growing Deep Learning team in Herzliya, working with physical, real-world signal data.

Description
In this role, you will help create the full data lifecycle that underpins our deep learning models: from designing what physical signal data we collect, through curation and quality monitoring, to model training and rigorous experimentation that drive model improvements. You will work closely with other Machine Learning and Data Engineering teams to ensure our models are trained on the best possible physical signal data, reaching the best accuracy, and that we deeply understand when and why they don't perform as expected - including building the tools to explain and interpret their behavior.

Responsibilities
Investigate model failures - identify patterns, hypothesize root causes, and work with the team to implement fixes, using explainability and interpretability methods where relevant
Own data curation: evaluate, clean, and curate physical signal datasets to maximize model training quality
Train, evaluate, and iterate on deep learning models using curated physical signal data, optimizing for accuracy, robustness, and generalization
Design and execute experiments end-to-end: from defining the question and data collection strategy, through analysis and statistical validation, to presenting clear conclusions and driving implementation
Define data collection strategies - collaborate with others to decide what data we should be collecting and why
Design and maintain monitoring solutions with others to ensure ongoing data quality and integrity at scale

Minimum qualifications
Ms.c. in Computer Science, Electrical Engineering, Computational Biology/Neuroscience, Mathematics, Statistics, or a related field
5+ years of industry experience in applied deep learning, data science, or a related field
Hands-on experience working with physical/real-world signal data
Strong hands-on experience with Python, PyTorch and SQL for large-scale signal/waveform data analysis and pipeline development
Hands-on experience with the full deep learning experimentation cycle: problem definition, data collection, statistical analysis, and conclusion-driven iteration
Proven ability to analyze model failures and translate findings into concrete improvements
Strong analytical thinking and ability to independently define and drive research directions
Excellent cross-functional communication skills - ability to work effectively with other Deep Learning and Data Engineers

Preferred qualifications
Experience with model explainability and interpretability methods - a strong advantage
Experience with continual or online learning - a strong advantage
Experience with data-efficient training strategies - a strong advantage
Experience with applied speech, audio, or signal processing deep learning systems
Familiarity with data quality frameworks, monitoring pipelines, and data validation at scale
Strong statistical foundation - hypothesis testing, uncertainty quantification, evaluation metrics design
Ph.D. in Computer Science, Electrical Engineering, Computational Biology/Neuroscience, Mathematics, Statistics, or a related field