Multi-Sigma-Alchemy is an integrated AI platform designed to accelerate materials and chemical R&D. Researchers can build AI models from their own experimental data for property prediction and experimental optimization, while utilizing diverse scientific data including molecular and crystal structures, spectra, and images. The platform also provides pretrained AI models for property prediction and enables candidate exploration using more than 22 million molecules with AI-predicted properties. Through practical materials R&D applications, this session will demonstrate how Multi-Sigma-Alchemy integrates prediction, discovery, and optimization into a unified AI workflow.