Summary
Saudi AZM is hiring an AI Engineer in Riyadh to build and operate production LLM systems, including fine-tuning, evaluation frameworks, multi-agent applications, and RAG solutions. The role also covers cloud and MLOps infrastructure, security, observability, disaster recovery, and CI/CD release practices. The engineer will lead and mentor a cloud/AI operations team and establish engineering standards. A bachelor's degree and 6–8+ years of relevant experience, including at least 2 years in applied AI or ML engineering, are required.
The posting
As published on azm.com. Pulled from the employer's site and shown as they wrote it.
AI systems
Build, fine-tune, and evaluate LLM systems for domain-specific tasks (QLoRA / PEFT on open-weight models such as Llama-3 and Mistral).
Design reproducible evaluation harnesses and A/B test frameworks with tracked metrics: task success rate, safety rate, and latency distributions (p50/p95).
Architect multi-agent and RAG systems (LangGraph, FastAPI, vector databases) from prototype through production.
Implement safety guardrails — input/output validation, allowlist/denylist policies, and controls that reduce invalid or high-risk model actions.
Translate business use cases into deployable prototypes with measurable acceptance criteria, and demo them to stakeholders.
Platform & infrastructure
Design and operate cloud infrastructure and MLOps workspaces (Azure, OCI, or GCP) for AI workloads on Kubernetes and containerized runtimes.
Build CI/CD pipelines and GitOps-based release promotion (Argo CD) across development, test, and production environments.
Implement end-to-end observability (Azure Monitor, Application Insights, ELK) with defined detection and response targets.
Apply network and perimeter security baselines (FW/WAF), automated code quality and SCA scanning (SonarQube, Black Duck), and gated pipelines.
Own disaster recovery design — automated backups, failover, and documented RTO/RPO commitments.
Engineering leadership
Lead and mentor a cloud/AI operations team; define monitoring, incident response, and release governance practices with clear uptime and MTTR targets.
Standardize SDLC practices — branching strategy, PR governance, release management, delivery reporting — to improve lead time and deployment frequency.
Consolidate engineering tooling and workflows; drive migrations and platform standardization where fragmentation slows delivery.
Produce handover documentation and runbooks that make systems auditable and operationally transferable.
Support vendor and licensing negotiations for cloud enterprise agreements.
Required Qualifications
Bachelor's degree in Software Engineering, Computer Science, or a related field.
6–8+ years in software, DevOps, or platform engineering, including at least 2 years in an applied AI or ML engineering capacity.
Proven delivery of production AI/LLM systems — not only research or notebook-stage work.
Strong Python; comfortable with Bash and YAML.
Deep hands-on experience with Kubernetes, Docker/Podman, and Terraform.
Production experience with at least one major cloud (Azure preferred; OCI or GCP acceptable).
Demonstrated ownership of CI/CD at scale (Azure DevOps, GitHub Actions) and GitOps release models.
Experience leading a team and setting engineering standards across multiple squads.
Preferred Qualifications
Master's degree in Applied AI, Machine Learning, or a related discipline.
Fine-tuning experience with QLoRA/LoRA on GPU clusters; PyTorch and Transformers.
Vector database experience (Milvus, Pinecone, or Weaviate) and RAG retrieval design.
Experience delivering on Saudi government or large-scale national digital platforms, with familiarity in local compliance and standards.
Arabic and English professional proficiency.
Technical Environment
Python, FastAPI, PyTorch, Transformers, LangGraph, Milvus/Pinecone/Weaviate, Redis, PostgreSQL, Kubernetes, Docker/Podman, Terraform, Argo CD, Azure DevOps, GitHub Actions, Azure ML, Azure Monitor / Application Insights, ELK, SonarQube, Black Duck, Fortinet FW/WAF.