This podcast follows from the presentation at https://youtu.be/LxFUB3uvosw. In this episode, we're joined by Alexander Visheratin, AI researcher at Playground AI. In this talk, Alexander Visheratin explores NLLB-CLIP: the current state of multilingual multimodal models, the challenges faced in their development, and his experience creating one of the top CLIP-like models that supports 200 languages.