Awni Hannun

Awni Hannun researches machine-learning algorithms and applications. His public biography lists his current role at Anthropic, earlier research at Apple, Zoom, Facebook AI Research, and Baidu, and a computer-science doctorate at Stanford. At Apple, he co-created MLX, a framework designed for machine learning on Apple silicon.

MLX gives developers array operations and model-building tools suited to the shared-memory architecture of modern Macs. Its open-source examples connect the framework to practical machine-learning tasks, helping researchers explore models without starting from a hardware abstraction of their own.

For someone experimenting on a Mac, MLX is a route into local model development and learning how computation moves between the processor and graphics hardware. The framework documentation and examples are the best starting point; memory requirements and model size still determine what fits on a particular machine.

Radar follows Hannun for the connection between research ideas and accessible computing tools. His software and technical writing help explain why a model behaves the way it does, while MLX provides a concrete place to explore those ideas. His earlier MLX work is distinct from his current employer’s products.