AI-Powered outbound Call System Using Asterisk SIP with Custom Caller ID

We are developing a fully automated AI voice calling system based on Asterisk SIP, capable of using custom Caller ID and conducting two-way voice conversations without human agents. We’re seeking an experienced developer to build and integrate the middleware pipeline, enabling seamless, real-time interaction between telephony and AI components.

⸻

Tech Stack Overview
• Asterisk – SIP dialer (custom Caller ID, DTMF support, audio streaming)
• Airtable – Lead database (stores client data, call status, webhook responses)
• n8n – Automation orchestrator (polls Airtable, triggers Asterisk, sends SMS, updates logs)
• Middleware – Streams live audio and manages STT → NLP → TTS flow
• Whisper – Real-time speech-to-text (STT)
• GPT-4o – Natural language generation using dynamic scripts
• ElevenLabs – Text-to-speech (TTS) with high-quality voice
• Twilio / Vonage – Sends post-call SMS with secure verification links

Your Role & Responsibilities
• Develop middleware to:
• Stream audio from Asterisk to Whisper
• Process transcriptions through GPT-4o
• Send responses to ElevenLabs for real-time voice generation
• Ensure natural, low-latency voice conversations (full-duplex)
• Capture and handle DTMF input (e.g., press 1 = verify, press 2 = to speak to an agent)
• Integrate n8n workflows and webhook triggers (sync with Airtable)
• Help structure the system to be modular, scalable, and secure


Ideal Skills & Experience
• Strong hands-on experience with Asterisk (ARI/AGI, SIP, Caller ID spoofing)
• Deep understanding of AI voice workflows (STT → NLP → TTS)
• Experience with Whisper, GPT-4, ElevenLabs, or equivalent tools
• Familiarity with n8n, Airtable, Firebase, Twilio/Vonage
• Ability to optimize for low-latency streaming and call logic