AI Voice System Enhancement

Project Overview: We are [login to view URL], an AI voice agent platform helping businesses manage calls efficiently. Our AI interacts like a human to handle tasks such as booking appointments, answering queries, and managing missed calls. We are looking for an experienced AI Engineer to help us refine and implement key features to improve functionality and user experience. Tasks and Responsibilities: Latency Optimization: Reduce latency to 800ms or lower for smooth real-time conversations. Ambient Sound Integration: Implement natural and high-quality background sounds with options like coffee shop, call center, and office environments. Ensure these sounds are smooth, realistic, and don’t interfere with call clarity. End Call and Transfer Call Functionality: Develop functionality where the AI understands when to transfer or end a call dynamically, based on caller intent and conversation flow. Advanced Settings Implementation: Interruption Sensitivity:Create a slider (0-1 scale) in the frontend allowing users to adjust interruption sensitivity. Provide clear documentation in the UI explaining what the sensitivity setting does. Reminder Message Frequency:Allow users to set how often the AI should send reminder messages in case of silence during calls. End Call on Silence:Add a timer slider in settings where users can define how long the AI waits during silence before ending the call. Appointment Booking System: Integrate with [login to view URL] for seamless appointment scheduling. Modify the prompt system to make it dynamic and human-like while ensuring users and callers receive email confirmations once an appointment is booked. Set up two levels of prompts: Generic Prompts (dynamic): System-wide prompts adaptable to any situation. Base Prompts (user-accessible): Editable prompts in the user dashboard for personalized AI agent behavior. Conversation Analysis Dashboard: Build analytics for each call, including: Call Summary and Transcription: Display user and AI dialogue separately with detailed logs. Call Metrics:Call Status (e.g., Successful, Ended). User Sentiment (e.g., Positive, Neutral). Disconnection Reason (e.g., Agent hangup). End-to-End Latency (e.g., 1210ms). Dynamic Responses: Replace static AI phrases with dynamic and context-aware dialogue for a more natural experience. Call Audio Recording: Ensure each call includes an audio recording with transcription logs and conversational analysis. Expectations: The AI must sound human-like in all interactions. Appointment booking should be seamless, with minimal effort for the caller. Emails for both the caller and user must confirm appointment details after booking. All features must be intuitive and easy to use in the frontend dashboard. Note: An audio sample demonstrating the desired conversational quality is attached for reference. Qualifications: Expertise in conversational AI and NLP technologies. Experience with Deepgram, Telnyx, Pinecone, and RAG models. Strong understanding of real-time systems and low-latency optimization. Proficient in building scalable front-end and back-end systems. Knowledge of VoIP technology and API integrations.