Autonomous Voice AI & n8n Patient Booking Architecture
Production-grade dental pipeline in self-hosted n8n: decoupled into lead verification, automated custom Voice agent generation via Vapi API, CRM sync, and bounce handling.
Open workflown8n & Voice AI Automation Engineer | Production AI Agents & Custom API Pipelines
About
I architect production-grade AI automations, autonomous agent workflows, and communication pipelines for B2B startups and modern service businesses.
Rather than fragile linear scripts, I focus on decoupled, fault-tolerant micro-architectures built in self-hosted n8n, Python, and modern CPaaS/Voice platforms:
• Autonomous Voice AI: Ultra-low latency voice agents built on Retell AI and Vapi for inbound appointment booking, qualification, and automated CRM logging. • Multi-Step n8n Pipelines: High-volume webhook processing, API rate-limiting management, and bidirectional CRM syncing (HubSpot, GoHighLevel, Airtable). • Telephony & Infrastructure: Deep VoIP and Asterisk engineering foundation ensuring enterprise-level reliability, error-handling, and zero dropped customer states.
Portfolio & Live Demos: https://sandrokharabadze.com Available for remote B2B contracts and technical integration partnerships.
Work samples · reviewed by HireAutomators
Production-grade dental pipeline in self-hosted n8n: decoupled into lead verification, automated custom Voice agent generation via Vapi API, CRM sync, and bounce handling.
Open workflowEvent-driven n8n pipeline interfacing with WhatsApp Cloud API to handle inbound leads, conversational state tracking, and instant CRM sync without dropped webhooks.
Open workflowSub-second latency voice assistant built on Retell AI/Vapi for qualifying buyers/sellers, handling live interruptions, and booking appointments directly into Cal.com/GHL.
Open workflowSelf-hosted n8n system rotating 5 mailboxes, paired with a 24/7 webhook listener that intercepts bounces instantly and updates databases to protect domain reputation.
Open workflowEngineered Asterisk and Linux SIP routing splitting inbound/outbound servers to eliminate bottlenecking during extreme spikes (up to 20,000 calls/day).
Open workflowTools & skills