Voice-Enabled AI Assistant Development
Budget: $30 – $120 USD
I’m building an AI-powered virtual assistant and need an experienced developer to take it from concept to a usable prototype. The core requirement is clear: the assistant must work seamlessly through both voice and text interactions—think conversational chat window plus hands-free, wake-word-style voice.
My ideal flow is:
• Speech-to-text captures the user’s question
• An NLP engine interprets intent
• The assistant fetches or computes the answer
• Text output is displayed and, when appropriate, spoken back with natural TTS
You’re free to recommend the best stack, but I’m already exploring Python with either Rasa or Dialogflow for intent handling, and something lightweight such as React or Flutter for the front end. If you prefer alternatives (e.g., Amazon Lex, Microsoft LUIS, or a custom PyTorch model), make the case—it just has to deliver fast, accurate responses.
Deliverables
• End-to-end prototype supporting voice + text, runnable on a local machine
• Well-structured, documented code (Git repo) with setup instructions
• Simple demo interface showing the dual interaction modes
• Read-me outlining how to extend the assistant with new skills or connect it to an external API
Acceptance Criteria
• Latency under two seconds for both voice and text round-trips on a standard laptop
• Word error rate < 10 % on common queries in English
• No hard-coded API keys or secrets; use environment variables
• Code passes a short linting test and includes unit tests for the intent parser
I’m ready to begin immediately and can clarify use cases as we iterate. Let me know what NLP libraries, ASR/TTS services, and deployment approach you’d leverage, plus a realistic timeline for the first working demo.
My ideal flow is:
• Speech-to-text captures the user’s question
• An NLP engine interprets intent
• The assistant fetches or computes the answer
• Text output is displayed and, when appropriate, spoken back with natural TTS
You’re free to recommend the best stack, but I’m already exploring Python with either Rasa or Dialogflow for intent handling, and something lightweight such as React or Flutter for the front end. If you prefer alternatives (e.g., Amazon Lex, Microsoft LUIS, or a custom PyTorch model), make the case—it just has to deliver fast, accurate responses.
Deliverables
• End-to-end prototype supporting voice + text, runnable on a local machine
• Well-structured, documented code (Git repo) with setup instructions
• Simple demo interface showing the dual interaction modes
• Read-me outlining how to extend the assistant with new skills or connect it to an external API
Acceptance Criteria
• Latency under two seconds for both voice and text round-trips on a standard laptop
• Word error rate < 10 % on common queries in English
• No hard-coded API keys or secrets; use environment variables
• Code passes a short linting test and includes unit tests for the intent parser
I’m ready to begin immediately and can clarify use cases as we iterate. Let me know what NLP libraries, ASR/TTS services, and deployment approach you’d leverage, plus a realistic timeline for the first working demo.