Dmytro U.

Voice AI and Automations Architect

Top 1% AI Voice and Automation Architect. 15+ large automation systems, 30+ multilingual Voice AI Agents over last 3 years across multiple industries. I automate small businesses that rely heavy on conversational AI - Voice AI agents, SMS, emails, chatbots. Keeping everything together in CRM with clean reports. My clients come usually from home services, car rentals, real estate and similar business - they have solid business, good reviews but stuck in operations. My approach: 1. Architecture first. With AI, executing the actual automation is much easier now. The hard part is designing the system right from the start - that’s why I focus a lot on understanding your business and plan priorities with you. 2. Weekly deployments. I structure the project into weekly iterations so you get value as soon as possible and we get users feedback. 3. AI-native documentation. Everything I build is documented in a format your own AI assistant (Claude or ChatGPT) can read. I connect it directly to your assistant so your team can just ask it. Voice Agents for inbound and outbound systems. I build reliable, production ready agents: - Low latency, high accuracy - Highly personalized calls - pull data from your CRM/DB before call - Multilingual (English, Spanish, French, Arabic) - Multiple industries (Home Services, Automotive, Recruiting, Healthcare) - Connected to ecosystem(CRMs, Calendars, PBX/VOIP systems/SIP) - Qualifications, Objection Handling, Appointments, Call Transfers, SMS The stack I work with: - WebRTC/Pipecat/Livekit/Ultravox for complex agents - Retell, Elevenlabs, VAPI for simple agents - n8n, make, zapier for automations - MCP servers to connect system with Claude/GPT assistants - Claude Code for rapid development in Python and JS - Twilio, Telnyx, custom SIP trunks - GoHighLevel, Hubspot, Zoho, Pipedrive - Clay, Apollo, Instantly, Smartlead - Supabase, Airtable, Clickup I'm bringing my 14 years software engineer, devops, tech lead experience to build Voice agents in a proven, agile and modern approach: - Evaluations driven development. Building Voice Agents based on datasets of multiple scenarios on how real conversations might go(qualifications, obj. handling, appointments, etc) - Versioning agents and automations in Github - Fine tuning/dynamic flows for complex agents - Kubernetes for custom deployments and handling of 100+ calls concurrently Languages I worked with when building Voice Agents: - English - French(Quebec) - Spanish - Gulf Arabic - Dutch I have own toolset/frameworks to build and evaluate voice ai agents rapidly.