AI Engineer

InnovationTeam; Riyadh; On-site; SAR 200,000 - 420,000; Full time

Job summary

InnovationTeam in Saudi Arabia seeks an AI Engineer to develop, deploy, and operate AI/LLM models across public and sovereign cloud environments (GCP for public workloads, Humain for classified data). You will build and fine-tune LLM/ML models for Arabic NLP, document classification, vision/OCR, and AIOps use cases, and own the serving stack with CI/CD and monitoring for latency and GPU utilization.

Responsibilities include pre-deployment evaluation, accuracy baselines, and optimizing inference

Qualifications

• 5 years ML/AI engineering, in production LLM deployment.

Responsibilities

• Run pre-deployment evaluation: accuracy baselines, regression and safety testing.
• Optimize inference: quantization, batching, context sizing; justify GPU allocation.
• Deploy on Humain GPUaaS: Kubernetes, GPU partitioning, quotas, RBAC.
• Build equivalent workloads on GCP (Vertex AI, GKE).
• Own serving stack (vLLM/TGI), model versioning, CI/CD, monitoring for latency, tokens, GPU utilization, drift.
• Ensure AI models comply with ZATCA data sovereignty and SDAIA guidelines.

Skills

Python

PyTorch

Hugging Face

Arabic NLP

LLM deployment

Tools

Kubernetes (production)

GCP Vertex AI

vLLM/TGI

Job description

Looking for an AI Engineer to Develop, deploy, and operate AI/LLM models across Clinets dual environment — GCP for public-cloud workloads, Humain sovereign cloud for classified data.

Build and fine-tune LLM/ML models for Arabic NLP, document classification, vision/OCR, and AIOps use cases.

Run pre-deployment evaluation

Accuracy baselines, regression and safety testing; evidence to justify GPU allocation.

Optimize inference — quantization, batching, context sizing — against measured usage.

Deploy on Humain GPUaaS: Kubernetes, GPU partitioning on B300 nodes, quotas, RBAC.

Build equivalent workloads on GCP (Vertex AI, GKE) with classification-based routing.

Own serving stack (vLLM/TGI), model versioning, CI/CD, and monitoring for latency, tokens, GPU utilization, and drift.

Ensuring developed AI Models Complying with ZATCA data sovereignty and SDAIA requirements (AI Ethics, GenAI Guidelines, PDPL).

5 years ML/AI engineering, in production LLM deployment with knowledge in

Skills

• Python, PyTorch, Hugging Face
• Kubernetes in production; GPU-served inference
• GCP Vertex AI or any equivellent cloud