Lead AI Systems Architect & Automation Head

Job TitleLead AI Systems Architect & Full-Stack Automation Head (Salary in USD) Job DescriptionThe Mission:We are a highly profitable ($1.2M/year profit) US-based firm. We don’t hire for "tasks"—we hire Intelligence. We are looking for a Founding Engineer in Egypt to act as the sole architect of an ever-evolving ecosystem of AI-driven products. You will have a direct impact on reducing overhead and maximizing revenue through pure technical innovation. If you want to build the "Brain" of a multi-million dollar company, this is your seat.The Capabilities You Must Possess:We are looking for the top 0.1% of engineers in Egypt. You must be a "Generalist Specialist" capable of rapid prototyping and deployment across any domain we dream up: Autonomous Intelligence: Mastery of LLM Orchestration (GPT-4o/Claude), Vector Databases (Pinecone/Milvus), and RAG. You build systems that learn. Global Voice Infrastructure: Expertise in VoIP/SIP (Asterisk/Kamailio) and Real-time Audio ( TECHNICAL STACK REQUIREMENTS 1. The "Scraper & Lead Gen" Pillar (Data Acquisition) Advanced Browser Automation: Expert in Playwright or Puppeteer (Stealth-plugins are mandatory). Anti-Bot Evasion: Proven ability to bypass Cloudflare Turnstile, DataDome, and Akamai using JA3 fingerprinting, TLS header manipulation, and residential proxy rotation (Bright Data/Oxylabs). Asynchronous Processing: High-concurrency scraping using Python Asyncio or Go to pull thousands of leads without bottlenecking. 2. The "Voice & Dialing" Pillar (Infrastructure) VoIP/SIP Engineering: Deep experience with Asterisk, FreePBX, or Kamailio. Ability to manage SIP Trunks and DID (Phone Number) pools. Regulatory Compliance (US/Global): Knowledge of STIR/SHAKEN protocols and "A-Level Attestation" to prevent calls from being labeled "Scam Likely." Low-Latency Audio: Expertise in WebRTC and Vapi/Retell/Telnyx to keep AI voice response time under 500ms. • Server Management: Ability to deploy local or cloud-based "Gateway Servers" in specific regions to bypass regional AI dialing restrictions. 3. The "Training Brain" & RAG Pillar (Intelligence) Vector Databases: Implementation of Pinecone, Milvus, or Weaviate to store and retrieve successful sales call data. Multimodal RAG: Building systems that don't just "read" text but "analyze" audio files to identify winning sales patterns from top reps. Speech-to-Speech (S2S): Integration with GPT-4o Audio or ElevenLabs to create human-like voices with emotional intelligence (tone, laughter, pauses). Diarization & Transcription: Using OpenAI Whisper to split recordings into "Rep" vs. "Customer" for automated performance grading. 4. The "Automation & CRM" Pillar (Logic) Agentic Frameworks: Proficiency in LangGraph, CrewAI, or AutoGen to build "Agents" that can autonomously update the CRM, book appointments, and handle customer service. Airtable/CRM Mastery: Deep API integration with Airtable, HubSpot, or GoHighLevel using Webhooks and Python SDKs. Mobile App Integration: Ability to build a lightweight Flutter or React Native app interface for D2D guys to access the "Real-time Training Brain" in the field. TECHNICAL STACK REQUIREMENTS 1. The "Scraper & Lead Gen" Pillar (Data Acquisition) Advanced Browser Automation: Expert in Playwright or Puppeteer (Stealth-plugins are mandatory). Anti-Bot Evasion: Proven ability to bypass Cloudflare Turnstile, DataDome, and Akamai using JA3 fingerprinting, TLS header manipulation, and residential proxy rotation (Bright Data/Oxylabs). Asynchronous Processing: High-concurrency scraping using Python Asyncio or Go to pull thousands of leads without bottlenecking. 2. The "Voice & Dialing" Pillar (Infrastructure) VoIP/SIP Engineering: Deep experience with Asterisk, FreePBX, or Kamailio. Ability to manage SIP Trunks and DID (Phone Number) pools. Regulatory Compliance (US/Global): Knowledge of STIR/SHAKEN protocols and "A-Level Attestation" to prevent calls from being labeled "Scam Likely." Low-Latency Audio: Expertise in WebRTC and Vapi/Retell/Telnyx to keep AI voice response time under 500ms. • Server Management: Ability to deploy local or cloud-based "Gateway Servers" in specific regions to bypass regional AI dialing restrictions. 3. The "Training Brain" & RAG Pillar (Intelligence) Vector Databases: Implementation of Pinecone, Milvus, or Weaviate to store and retrieve successful sales call data. Multimodal RAG: Building systems that don't just "read" text but "analyze" audio files to identify winning sales patterns from top reps. Speech-to-Speech (S2S): Integration with GPT-4o Audio or ElevenLabs to create human-like voices with emotional intelligence (tone, laughter, pauses). Diarization & Transcription: Using OpenAI Whisper to split recordings into "Rep" vs. "Customer" for automated performance grading. 4. The "Automation & CRM" Pillar (Logic) Agentic Frameworks: Proficiency in LangGraph, CrewAI, or AutoGen to build "Agents" that can autonomously update the CRM, book appointments, and handle customer service. Airtable/CRM Mastery: Deep API integration with Airtable, HubSpot, or GoHighLevel using Webhooks and Python SDKs. Mobile App Integration: Ability to build a lightweight Flutter or React Native app interface for D2D guys to access the "Real-time Training Brain" in the field.