Donato Capitella

Donato Capitella creates content focused on running and optimizing local AI models, including large language models and coding agents, on various hardware platforms. The channel provides detailed benchmarks and configuration guides for systems using AMD Strix Halo, Radeon AI Pro cards, and other GPUs like the NVIDIA V100, with a focus on tools such as llama.cpp, vLLM, and Unsloth.

The videos cover technical processes like model fine-tuning, multi-token prediction for faster inference, and setting up distributed and multi-node AI systems. Donato Capitella describes being a tech enthusiast who likes to discover how things work by breaking them apart.