Pushing the Limits of On-Device Streaming ASR: A Compact, High-Accuracy English Model for Low-Latency Inference

Publish a highly optimized C++/Rust engine for running compact ASR models on CPU. The focus should be on sub-100ms latency on resource-constrained embedded systems.

Suggested repo: NanoSpeech

"State-of-the-art ASR on a CPU—no GPU required."

Estimated effort: 150h