oMLX Creator Jun Kim Joins Hugging Face to Advance MLX

Hugging Face has announced the hiring of Jun Kim, creator and lead maintainer of oMLX, to formally support the development and expansion of the MLX ecosystem. The appointment signals a strategic commitment by Hugging Face to local artificial intelligence and aims to provide institutional backing for open-source machine learning tools optimized for Apple Silicon. MLX, an open framework developed by Apple for efficient local AI inference, has rapidly become a cornerstone of the on-device machine learning landscape. Since its initial release in late 2023, the framework has gained significant traction across the developer community. Hugging Face has long operated as the primary hub for discovering and sharing MLX-compatible models, and this acquisition represents a deliberate effort to formalize ecosystem support. Under the new arrangement, oMLX will transition from a community-driven initiative to a fully funded and professionally maintained project. Kim will continue to lead the repository, which remains licensed under Apache 2.0. Increased corporate backing is expected to accelerate development cycles, improve long-term stability, and establish clearer architectural guidelines for external contributors. Strategically, oMLX will function as a testbed for experimental features while leveraging established foundational libraries such as mlx-lm and mlx-vlm. Hugging Face has identified a primary technical objective: streamlining the conversion of models defined in the Hugging Face Transformers library into reference MLX implementations. By standardizing these conversion pipelines, the company aims to allow multiple inference engines to build upon shared model definitions, enabling each platform to focus on specialized optimizations and unique capabilities. The initiative also seeks to deepen collaboration with core MLX stakeholders, including the teams behind mlx-lm, mlx-vlm, and LMStudio. Hugging Face has stated that successful upstream contributions will be directed to the most appropriate open-source repositories, fostering a more interoperable and efficient development environment. By integrating oMLX into its operational structure, Hugging Face aims to reduce friction for developers deploying local AI workloads. The move aligns with broader industry shifts toward decentralized computing and on-device intelligence, particularly as Apple Silicon continues to expand its market share in consumer AI hardware. With dedicated engineering resources and centralized coordination, the MLX ecosystem is positioned to mature rapidly, delivering more robust infrastructure for researchers and engineers building next-generation local AI applications.