We’re Yokeru (yokeru.ai). Our AI voice agents make welfare calls for care providers. When an elderly person doesn’t answer, our integrations make sure a human follows up. The plumbing has to work.
We need someone who builds this stuff already:
- Webhooks and serverless functions (AWS Lambda, n8n)
- REST APIs into CRMs and alarm platforms
- Retell or VAPI agent config and prompt work
You’ll own integrations end to end, from prompt to CRM record.
Contract, remote, $20-50/hr depending on experience. UK-timezone overlap needed. We start with a short paid trial project, then ongoing work with a small team that ships fast.
Apply here: https://integrations-apply.vercel.app/
Takes 10 minutes. Show us something you’ve built with Retell, VAPI, or n8n.
2 curtidas
Hi there! This role fits my profile perfectly. My core expertise is n8n workflow development and building the “behind-the-scenes plumbing”—connecting APIs, Webhooks, and syncing CRM data flawlessly so the team can focus on growth.I am currently away from my development PC and can’t record the Loom video right this second, but I don’t want to miss out on this. I’m sending an email to your team right now to introduce myself and will submit the official form with the video as soon as I’m back at my desk. Looking forward to connecting!
Hi — I build production n8n workflows and integrations (webhooks, REST APIs into CRMs, Retell/VAPI voice agent configs) using Claude Code as part of my dev workflow, so I move fast on this kind of plumbing. I currently run 255+ workflows in production on my own n8n instance, so I know the platform well. Happy to share a recent build. What’s the best way to send it over — here or the apply link?
Hi Monty, your Yokeru setup — AI voice agents flagging missed welfare calls and routing follow-up to a human via n8n — is exactly the kind of voice-plus-automation pipeline I build daily for my own real estate and e-commerce operations. The plumbing challenge you mention (Retell/VAPI event → webhook → CRM update, reliably, with retries) is where most of these builds break in production. First step I’d take: map your current call-to-CRM path and pressure-test the webhook handling for dropped events before touching prompts. Happy to build a small working demo of that flow so you can see it firsthand before any commitment. DM me here or check Abderrahmane | Profile | Fiverr.
Hi Monty — shipping this last week made me think of your welfare-call flow.
Retell voice agent for US healthcare appointment reminders — same shape as an eldercare welfare check minus the alarm-platform escalation: intent-confidence gating on ambiguous “maybe?” responses (they don’t get logged as confirmations), Twilio SMS fallback when voice drops, hash-chain audit trail for the compliance folks. Voice Agent — Healthcare Appointment Reminder + Callback
Different domain — but the “did they actually understand?” problem rhymes, and the failure modes are worse on your side (loneliness signal vs a missed appointment slot). One design question I’d want to test against Careium/Appello escalation contracts: how do you handle low-confidence intent when the escalation contract requires a definitive answer? On healthcare it just routes to human review — welfare probably needs a tighter contract.
Would trade notes on the intent-register mapping if useful.
— Priyanshu Kumar
@Montyyokeru
hey! intent confidence is the tricky part with voice AI agents. if the confidence score is low, you can’t let the model guess — it has to branch straight to an SMS or human fallback.
in n8n i usually handle this by having the LLM output confidence scores in structured JSON, routing low confidence runs directly to a Twilio SMS module, and logging every execution step to a SQL database for HIPAA compliance.
available for UK hours and remote work. sent you a DM to check out your current voice setup.