38 jobs for "AIML - Data Scientist, Evaluation"

The Personalized AI for Science (PAIS) team is at the forefront of machine learning and health science at Apple. We are a close-knit team of deeply technical AI/ML research scientists who incubate ambitious research ideas and develop new statistical/ML methodologies, ultimately in service of new Apple health and wellbeing applications. We are seeking a research scientist who combines deep learning (and/or signal processing) experience with a rigorous foundation in statistics. The ideal candidate is passionate about model robustness and scientific rigor, ensuring that the models we deploy meet the high evidentiary standards required of health applications.

Description
In this role, you will design, train, and adapt modern deep learning models on unique health data, and apply your statistical expertise to evaluate them rigorously - going beyond aggregate accuracy to interrogate model behavior across populations, distributional shifts, and clinically meaningful subgroups. You will be expected to engage substantively with health as an application area, developing the domain knowledge needed to ask the right modeling questions and evaluate models against the realities of clinical and wellbeing use. You will work closely with collaborators across research and product teams, and contribute to features that ultimately reach a billion Apple devices worldwide.","responsibilities":"Designing, training, and adapting modern deep learning models and architectures on large-scale physiological and behavioral data
Designing rigorous evaluation protocols, including subgroup analyses, robustness to distributional shifts, and characterization of failure modes
Applying statistical methods (hypothesis testing, uncertainty quantification, causal inference, survival analysis) to draw defensible conclusions from observational health data
Translating clinical and product questions into well-posed problem definitions
Engaging with the academic community by publishing papers and presenting your work
Collaborating with product teams across Apple to transition research into deployed health features

Preferred Qualifications
Proficiency in Python and modern ML/analysis stacks, including LLM-based coding and analysis
Excellent communication skills, including the ability to present technical work to clinical and broad technical audiences
Research or applied experience in deep learning and statistical modeling in the health domain (e.g. wearable devices, clinical studies, or electronic health records)
Experience with one or more of: causal inference, survival analysis, time-series modeling, or longitudinal analysis
Experience with rigorous evaluation methodology for ML in clinical or scientific settings, such as benchmarking under controlled distribution shifts, calibration analysis, or regulatory-grade validation
Publications at venues such as NeurIPS, ICML, ICLR

Minimum Qualifications
Hands-on experience designing and training modern deep learning models and architectures, including foundation models, self-supervised pretraining, multimodality, and parameter-efficient adaptation
Experience characterizing and improving the robustness of ML models under distributional shifts that are common in health data (subject variability, device heterogeneity, temporal drift)
Strong foundation in statistics, with the ability to draw defensible conclusions from observational data
Ability to design and execute non-trivial model evaluations beyond standard validation

About Apple

Working at Apple means doing more than you ever thought possible and having more impact than you ever imagined.

Size: 10,000+ employees
Industry: Technology

View company profile