Voice conversations between Gemma4 12B and E2B on GPU and Jetson Orin

Voice conversations between Gemma 4 12B and E2B are demonstrated on an RTX PRO 4500 Blackwell GPU and a Jetson Orin NX 16GB, respectively. Both setups utilize a reSpeaker Flex 4-mic array and a 3W speaker. Inference is powered by Cortexist Little Gemma, a C-based LLM engine for CUDA devices, which outperforms llama.cpp on Jetson Orin and maintains performance with long voice prompts. The open-source pipeline also supports lip sync, expressions, and gestures.