Founding Engineer, Conversational Voice AI 🎙️

SayVo AI | Full-time | Remote (North America) | $35–50/hr CAD (~$73K–$104K/yr) + equity for the right person

SayVo builds AI voice agents for companies across North America. Our agents call new leads in under 60 seconds, qualify them, handle objections, and book appointments straight into the client's CRM. It's the work that used to take a room full of cold callers.

We're building our own conversational voice AI platform, and we're hiring the engineer who builds it.

This is a founding engineering role. You'll own our core voice infrastructure end to end, ship it into live client accounts, and grow into the technical leader of the company as we scale.

WHAT YOU'LL BUILD

✅ A real-time conversational voice pipeline: streaming speech-to-text, LLM orchestration, and speech synthesis fast enough to feel like a natural phone call

✅ Turn detection that holds up on real calls. VAD, endpointing, and model-based prediction of when a caller has actually finished speaking rather than just paused

✅ Barge-in and interruption handling, with echo cancellation clean enough that the agent never talks over itself

✅ Latency engineering against real budgets: sub-500ms streaming transcription, sub-200ms time to first audio, and a total response time that keeps the conversation natural

✅ Telephony and transport: SIP, WebRTC, DTMF, call transfer, voicemail detection, and the realities of live PSTN traffic

✅ Concurrency and reliability at scale. Hundreds of simultaneous calls without latency degrading

✅ Conversation state, function calling, and mid-call CRM writes that cannot fail silently

✅ Observability: recordings, transcripts, traces, and evals, so agent quality is measured rather than guessed

YOU'LL ALSO

✅ Ship client builds. Voice agents wired into CRMs like GoHighLevel, Pipedrive, and Bonzo using n8n, webhooks, and APIs

✅ Own the telephony and compliance layer: Twilio numbers, A2P/10DLC registration, call and SMS flows

✅ Write and tune the prompts and scripts that make our agents sound like a top human rep, not a bot

✅ Build the internal software that automates our client onboarding

✅ Work with clients directly. Run setup calls, pull what you need from them, and explain technical things in plain English

✅ Contribute to growth: email campaigns, organic content on our company LinkedIn, paid creative

We're an early-stage team. Roles stay fluid, and the person in this seat wears more than one hat.

THIS ROLE IS FOR YOU IF

• You've built real-time, stateful systems before and you know why they're harder than they look
• You're obsessive about detail. You double-check the field mapping because you know exactly what breaks if you don't
• You can read an API doc and a JSON payload, trace why a webhook is failing, and fix it without hand-holding
• You measure before you optimize, and you can defend a latency budget with data
• You can join a client call and explain something technical clearly and with confidence
• Social intelligence: you understand how people think, and you can write AI that sounds natural and persuasive rather than scripted
• You want one role to pour yourself into and get genuinely great at
• You want real ownership of a product, not a queue of tickets

THIS ROLE IS NOT FOR YOU IF

• You're juggling other clients, courses, a side business, or another job. This is full-time and single-focus
• You're looking for a low-intensity remote role. We move fast, ship weekly, and the pace is demanding
• You need step-by-step direction. We're building the playbook, not following one
• You only want the interesting half of the work
• "Close enough" is anywhere in your vocabulary. When something misfires, a real client's leads don't get called that day
• You'd rather stay behind a screen than talk to a client

WHAT WE NEED (HARD REQUIREMENTS)

• You can architect and ship a production backend independently. Language matters less than judgment, though most of this work runs on Python or TypeScript
• Real experience with real-time systems: websockets, streaming, async, event loops, and the failure modes that only appear under load
• A working understanding of the modern voice stack: streaming STT, TTS, LLM orchestration, and where latency actually accumulates
• Strong with APIs, webhooks, and JSON debugging
• A sharp, clear communicator we can put in front of a client with confidence
• You own delivery end to end and stay composed under pressure
• Highly responsive and flexible with your hours. When a launch needs you, you show up
• Full-time, based in North America (US/Canada), working overlapping hours

BONUS POINTS

• You've shipped a production voice agent, on your own stack or on a platform
• Hands-on with LiveKit Agents, Pipecat, or a comparable orchestration framework
• Telephony experience: Twilio, Telnyx, SIP trunking, carrier behaviour
• Production work with Deepgram, AssemblyAI, Cartesia, ElevenLabs, or Rime
• Prompt engineering for conversational or sales AI
• You know the mortgage / lending world, or you've done real sales or cold calling
• Automation platform experience (n8n, Make, or similar)

WHAT'S IN IT FOR YOU

• $35–50/hr CAD, full-time (about $73K–$104K/year), based on what you bring
• Equity / options on the table for an exceptional hire
• Direct access to the founder, real ownership, and a clear path to technical leadership as we grow
• You own the most important system in the company from day one
• A front-row seat in one of the most exciting corners of AI. We're automating an entire job category, and you'd be building the engine

HOW TO APPLY

Hit Apply on LinkedIn.

Optional, and it earns extra consideration: send a 2 to 3 minute video introduction to arman@sayvo.ai covering who you are and how you'd approach building a platform like this. Keep it simple, we're not judging production quality.

We read every application that follows these instructions.