F2LLM-v2: Inclusive, Performant, and Efficient Embeddings for a Multilingual World | Signal Canvas | ScienceToStartup