Daniel Han

Daniel Han is the Co-founder and CEO of Unsloth AI, an open-source artificial intelligence startup based in Sydney, Australia. Previously, he served as a software engineer at NVIDIA Corporation, where he optimized core machine learning algorithms, including accelerating t-SNE performance by 2,000 times. Han also created and currently maintains Hyperlearn, an open-source machine learning package utilized by engineers at organizations such as NASA and Microsoft. As an active contributor to the broader AI ecosystem, he has identified and resolved over 20 core bugs in prominent foundational models, including Meta's open-source Llama. In early 2024, he led Unsloth AI through the Y Combinator Winter 2024 batch to further develop tools for memory-efficient model training. His current professional focus remains on building infrastructure that accelerates the fine-tuning of large language models for the global developer community.