An ultra-efficient 1.7B sovereign assistant, built on SmolLM2 and refined via SFT to bridge the gap between low-latency local inference and high-fidel
Chat with Cygnus Alpha AI assistant