MLOps/LLMOps Engineer at Social Health Insurance Fund

MLOps/LLMOps Engineer
Social Health Insurance Fund
1,500,000 to 2,000,000 KZT Gross per month
Astana, office

The Social Health Insurance Fund finances healthcare services under Kazakhstan’s Guaranteed Benefits Package and Mandatory Social Health Insurance system. It is expanding its digital services, analytics capabilities, and artificial intelligence initiatives. The Fund is building an Artificial Intelligence function and secure AI solutions to support planning, control processes, and healthcare delivery. Its initiatives include internal AI assistants and retrieval-augmented generation (RAG) services.

Responsibilities:
• Design and maintain secure infrastructure for ML and LLM products, including access controls, audit logging, secrets management, and safeguards for personal and medical data.
• Build end-to-end MLOps pipelines for data preparation, training, validation, deployment, monitoring, and retraining.
• Deploy and optimize LLM services, including local models, inference servers, APIs, GPU infrastructure, and performance monitoring.
• Develop RAG solutions for internal documents, regulations, and knowledge bases; manage vector databases, document indexing, search, and data refresh processes.
• Establish CI/CD workflows for models, prompts, agentic workflows, and AI services.
• Monitor model quality, latency, errors, inference costs, and compute utilization.
• Collaborate with Data Scientists, ML Engineers, Data Engineers, analysts, and business teams to deliver AI use cases.
• Contribute to technical stack selection and define MLOps/LLMOps standards.

Requirements:
• 2–3+ years of experience in MLOps, LLMOps, DevOps, ML Engineering, or Data Engineering.
• Strong Python skills and experience building APIs with FastAPI or similar frameworks.
• Hands-on experience with Docker, Kubernetes, Git, and CI/CD.
• Solid understanding of the ML lifecycle, including data preparation, training, evaluation, deployment, monitoring, and retraining.
• Experience with Airflow, MLflow, S3/MinIO-compatible object storage, PostgreSQL, and/or ClickHouse.
• Practical knowledge of Hugging Face, Transformers, vLLM, RAG, embeddings, and vector databases; experience with one or more of Qdrant, pgvector, Milvus, Weaviate, Elasticsearch, or OpenSearch.
• Understanding of agentic workflows, tool calling, LLM orchestration, and LLM evaluation.
• Experience with observability tools such as Prometheus, Grafana, OpenTelemetry, ELK/EFK, Langfuse, LangSmith, or similar platforms.
• Understanding of IAM, secrets management, security logging, and audit requirements.
• Ability to design reliable systems, document technical decisions, and communicate effectively with non-technical stakeholders.

Optional:
• Experience serving models on GPU infrastructure.
• Experience with Terraform, Helm, Argo CD, GitLab CI, or GitHub Actions.
• Experience with AWS, Azure, GCP, private cloud, or on-premises environments.
• Experience in finance, government, healthcare, insurance, or other regulated industries.
• Experience with OCR, document classification, entity extraction, semantic search, or RAG.
• Experience with fraud detection, forecasting, anomaly detection, or risk-scoring models.