Senior Solo AI Engineer (Python, RAG, Conversational and Voice AI)

Role Overview

What You Will Build

• Develop multi-step chat journeys with intent detection, tool calling, confirmations, and transaction records.
• Build streaming voice systems with speech recognition, speech synthesis, interruption handling, and visual handoff.
• Create RAG pipelines for PDF, DOCX, scanned, and bilingual documents with retrieval evaluation.
• Develop agentic workflows with deterministic controls to prevent invented fees, rules, or records.
• Serve hosted and local open models, including GPU tuning, quantization, latency, and throughput optimization.
• Integrate SOAP, WSDL, XML, REST, UAE PASS, OAuth, payment gateways, callbacks, receipts, and audit logs.
• Deploy with containers, systemd, reverse proxies, and TLS, including complete offline packages for air-gapped sites.

Working Environment

• The role is primarily on-site and may involve work in air-gapped government networks.
• The engineer may be the only engineer on a delivery and will work directly from client documents and existing code.

Required Experience

• At least 4 years of production Python experience with asynchronous services, APIs, background workers, and error handling.
• Production experience shipping LLM applications using RAG, tool calling, or agents.
• Experience with hosted model APIs and either open-model GPU serving or the ability to learn it quickly.
• Strong Linux skills covering containers, networking, TLS, and offline server debugging.
• Ability to implement and diagnose integrations from WSDL or API specifications.
• Independent working style and accountability for end-to-end delivery outcomes.

Strong Pluses

• Arabic, especially Gulf dialect, and English-Arabic code-switching.
• Streaming speech systems such as ASR, TTS, WebSockets, or WebRTC.
• GPU serving and quantization with vLLM, TensorRT-LLM, or similar tools.
• UAE PASS, UAE government services, or payment gateway integrations.
• Air-gapped packaging, offline repositories, checksums, SBOMs, or SQL Server.

How the Team Works

Requirements

• At least 4 years of production Python experience, including asynchronous services, APIs, background workers, and real error handling.
• Production experience shipping LLM applications using RAG, tool calling, or agents.
• Experience with hosted model APIs and either running open models on GPUs or the ability to learn GPU model serving quickly.
• Strong Linux skills, including containers, networking, TLS, and debugging servers without internet access.
• Ability to implement and troubleshoot integrations from WSDL or API specifications.
• Independent end-to-end working style with ownership of delivery outcomes.
• Arabic, especially Gulf dialect, and English-Arabic code-switching experience.
• Experience with streaming speech systems, including ASR, TTS, WebSockets, or WebRTC.
• Experience with GPU serving and quantization using vLLM, TensorRT-LLM, or similar tools.
• Experience with UAE PASS, UAE government services, or payment gateway integrations.
• Experience with air-gapped packaging, offline repositories, checksums, or SBOMs.
• Experience with SQL Server or comparable enterprise databases.