Apple's Data Analytic and Quality (DAQ) group is seeking a Machine Learning/Data Scientist to specialize in the evaluation of GenAI models.
Requirements
- Strong proficiency in Python and experience with ML frameworks such as TensorFlow and PyTorch
- Solid understanding of computer vision principles and extensive experience with 3D deep learning techniques, including NeRF, PointNet, MeshCNN, and Diffusion Models
- Hands-on experience with image/video/3D generative ML models and a proven track record of working with 3D data (point clouds, meshes, voxels, signed distance functions)
- Proficiency in 3D data processing and rendering frameworks such as Blender API, Unreal Engine, Open3D, Kaolin, or Minkowski Engine
- Experience with real-time rendering and physics-based simulations
Responsibilities
- developing methods for the evaluation and improvement of foundation models
- refining the data used in training the models
- design and develop image/video analysis (ROI, SOD, content classification, artifact detection, etc.) and quality assessment algorithms
- crafting testing strategies and solutions for E2E quality evaluation with a combination of image/video analysis, processing, and encoding algorithms
- conduct detailed failure analysis on large GenAI models and machine learning models in different aspects like Operational/Ethical/Security & Privacy
- Work directly with developer teams to define and request data to ensure adequate data coverage
- Prototype novel evaluation and benchmark methods from foundation model literature research
Other
- BS and a minimum of 3 years relevant proven experience
- collaboration with innovative teams at Apple focused on developing foundation models, including ML engineers, Data Scientists, and ML Infrastructure Engineers
- Develop tools for analyzing and visualizing data
- Design and implement experiments (DOE) for engineering studies and large-scale user studies and analyze results
- Facilitate and support data collection and contribute in defining feature specifications and anticipated user experience based on data insights
