Frontier ML Efficiency Research Scientist

Apple Inc.; Zürich; Vor Ort; CHF 140.000 - 210.000; Vollzeit

Zusammenfassung

Apple Inc. in Zürich seeks research scientists and engineers to advance machine learning, focusing on frontier models and efficiency of LLMs and diffusion models. You will join a team of world-class researchers powering Apple technologies and contributing to publications.

The role emphasizes designing, training, and evaluating models, collaborating with product teams, and publishing findings. Strong backgrounds in ML, Python, and modern frameworks are expected.

Qualifikationen

• Significant prototypes delivering state-of-the-art performance in ML domains.
• Publication record at top conferences (e.g., NeurIPS, ICML, ICLR).
• Experience across ML stack from systems to algorithms.
• Fluency in Python and solid software engineering.
• Expertise in ML algorithms and best practices for deep learning.
• Proficiency with ML modeling frameworks such as PyTorch and JAX.

Aufgaben

• Designing, training, and adapting frontier models to improve efficiency and performance.
• Designing and conducting empirical studies to understand frontier models.
• Developing prototypes to demonstrate model and algorithm effectiveness.
• Engaging with the academic community by publishing papers and presenting work.
• Collaborating with product teams across Apple to ship research.

Kenntnisse

Prototype dev

SOTA ML

Python fluency

ML algorithms

Research publication track record

Cross‑domain collaboration

Ausbildung

Bachelor/Master/PhD in CS/EE/ME or related

Tools

PyTorch

JAX

Jobbeschreibung

Apple Inc. in Zürich seeks research scientists and engineers to advance machine learning, focusing on frontier models and efficiency of LLMs and diffusion models. You will join a team of world-class researchers powering Apple technologies and contributing to publications.

The role emphasizes designing, training, and evaluating models, collaborating with product teams, and publishing findings. Strong backgrounds in ML, Python, and modern frameworks are expected.