AI Architect

Position overview

Location and work mode

• Based 100% onsite in Abu Dhabi, UAE.
• Immediate or short-notice joining is preferred.

Employment and compensation

• Full-time, long-term employment.
• Monthly salary is AED 20,000–22,000, depending on skills and experience.
• Benefits include UAE employment visa subject to requirements, medical insurance, applicable air ticket, and other standard employment benefits.

Key responsibilities

• Design conversational AI, agentic AI, and enterprise LLM solutions; develop voice agents, chatbots, and real-time STT/TTS systems.
• Build RAG pipelines and orchestration; deploy and optimize inference using vLLM, Hugging Face TGI, and TensorRT-LLM.
• Tune PagedAttention, KV cache, batching, GPU utilization, quantization, and distributed inference.
• Architect NVIDIA GPU infrastructure and deploy workloads on Kubernetes using GPU Operator and MIG.
• Support bare-metal clusters, scaling, production reliability, monitoring, observability, and troubleshooting.
• Collaborate with engineering and business teams to deliver secure, reliable production AI solutions.

Technical requirements

• Mandatory expertise includes conversational and agentic AI, voice agents, chatbots, STT/TTS, RAG, LLM orchestration, and production deployments.
• Required inference technologies include vLLM, Hugging Face TGI, TensorRT-LLM, PagedAttention, KV cache, continuous batching, FP8, INT4, AWQ, and GPTQ.
• Required GPU expertise includes NVIDIA H100, H200, and A100, distributed inference, tensor and pipeline parallelism, and GPU tuning.
• Required platform and monitoring skills include CUDA, Triton, FlashAttention, Kubernetes, NVIDIA GPU Operator, MIG, bare-metal clusters, Prometheus, Grafana, and DCGM.

Candidate eligibility

• At least five years of relevant hands-on technical experience and demonstrable production deployment and operation of AI/LLM solutions.
• Strong understanding of GPU-accelerated inference and performance optimization.
• Ten or more years is preferred for Senior/Lead Architect candidates.
• Candidates based in the UAE or India and immediate or short-notice joiners are preferred.
• Must meet applicable UAE work authorization, visa, and client deployment requirements.

Candidate submission and interview

• Submissions should include an updated CV highlighting AI/LLM implementations, current location, onsite willingness, salary details, notice period, and a summary of production experience.
• Include relevant projects involving vLLM, TensorRT-LLM, GPU optimization, or AI voice agents.
• The process includes technical screening and a detailed client interview; same-day technical interviews may be arranged.