Description
We are looking for an experienced speech recognition / ASR engineer to improve the recognition of spoken German number words in our Windows educational application. What We Need: Analyze the current implementation and improve the speech recognition. We are open to the best solution, for example: * Optimized Vosk * Constrained grammar * Streaming ASR * Sherpa-ONNX * Keyword spotting * Any better local ASR approach Requirements: * Experience with speech recognition / ASR * Experience with Vosk, Kaldi, Sherpa-ONNX, or similar * German ASR experience * Experience with limited-vocabulary recognition is a plus When Applying please include: 1. Relevant ASR experience. 2. Similar projects. 3. Your recommended technical approach. 4. Estimated time and cost. We are looking for an expert who can solve this specific speech recognition problem, not rebuild the entire application.