تخطي إلى المحتوى الرئيسي
Adel Atya

Adel Atya

Voice AI & Forward Deployed Engineer - production agents on Retell and Vapi

كفاءة
Cairoعضو منذ ١٤ أكتوبر ٢٠٢٦
VapiTwilioconversation designtool callingPythonFastAPINodeRESTwebhooksRenderGoHighLeveln8nLangGraph
·

وظّف هذا الشخص

أرسل عرض عمل مباشرة لهذا المرشح

VapiTwilioconversation designtool callingPython

نبذة

I build production voice agents alongside the businesses that use them. I design the conversation and the escalation rules, wire them into the customer's CRM and calendar, then debug what breaks in production - webhooks delivered twice, a tool call that times out mid-sentence, an agent that interrupts a caller who was still talking. Seven agents are live and dialable right now, one per vertical.

الخبرة

Founder & AI Voice Agent Engineer at adelatya.com (2025 – Present)

  • Design, deploy and support production voice agents for US and Gulf service businesses on Retell and Vapi over Twilio - conversation design, tool calling, CRM and calendar integration, human handoff rules, and a live operator dashboard (FastAPI + Postgres on Render) showing every call, transcript and outcome.
  • Onboard each customer end to end: pull their hours, services, pricing and objections; script the agent in their voice; run test calls with the owner; measure a before-and-after baseline of calls answered and jobs booked.
  • Operate the infrastructure around it - scheduled jobs, monitoring, failure alerting, automatic pause on deliverability or auth failure, and DNS/mail authentication (SPF, DKIM, DMARC, BIMI).

Software Developer, Integrations & Automation at theTrendUp (2024-01 – Present)

  • Three plugin suites across 30+ live business sites feeding one GoHighLevel CRM; self-provisioning tenant onboarding via the GHL REST API, idempotent and retry-safe, with replay queues so no lead is dropped.
  • Security-scoped REST bridge letting an AI agent operate a live production CMS - token auth, dual authorisation, mandatory-draft writes; rejected an endpoint that would have allowed arbitrary code execution.

AI Code Reviewer at Drytis (2026-09 – Present)

  • Vetted engineer on Drytis, where human engineers review, test and finish software produced by AI coding agents - codebases in Laravel/PHP, Next.js/TypeScript, Python and Node.
  • Test the failure modes automated suites miss: access-control and IDOR gaps, server-side validation bypassed by a disabled control, races between concurrent writes, state that must survive a reload, and layout behaviour at narrow widths. Deliver reproducible defect reports with fixes the customer can apply directly.

التعليم

BSc in Computer Science (Artificial Intelligence) – The British University in Egypt (2026)

التقييمات

ذات صلة