Senior AI Engineer – LLMOps & MLOps

Role Overview

Senior AI Engineer responsible for end-to-end ownership of AI initiatives for a claims processing domain. This is an execution-focused role in the company’s AI Transformation Office, focused on bridging legacy insurance data systems with modern cloud AI services.

Mission

• Build the automated infrastructure connecting legacy data systems to AWS and Azure AI services.
• Own the “Ops” of AI: deploy, observe, and scale LLM applications, RAG pipelines, and traditional ML models in a multi-cloud environment.

Key Responsibilities

• Multi-cloud pipeline execution: Build and maintain automated CI/CD and Continuous Training (CT) pipelines across AWS (SageMaker/Bedrock) and Azure (AI Studio).
• LLMOps / RAG infrastructure: Implement RAG infrastructure, including vector database management (OpenSearch, Pinecone, or Azure AI Search) and semantic index optimization.
• Legacy data connectivity: Create secure ingestion and data movement “pipes” from Mainframes, SQL Server, and other on-prem databases into cloud-native MLOps workflows.
• Automated model evaluation: Implement evaluation frameworks for LLMs (LLM-as-a-judge, ROUGE, METEOR) and validation for traditional ML before deployment.
• Observability & monitoring: Add real-time monitoring for model drift, hallucinations, latency, and token consumption to manage quality and cost.
• Infrastructure as Code: Manage AI resources with Terraform or CloudFormation, following Privacy by Design.
• Advanced analytics integration: Work with teams using Palantir, Databricks, or Snowflake to ensure high-fidelity data flow into production models.
• IT & security collaboration: Partner with IT/Security on IAM, VPC peering, and firewall configurations.
• Scalable inference engineering: Optimize serving endpoints for low latency/high throughput, using Docker/Kubernetes and serverless architectures as appropriate.
• Prompt & model versioning (PromptOps): Ensure auditability with rigorous version control for prompts, model weights, and data snapshots.
• Data science engineering enablement: Automate feature stores, feature engineering pipelines, and productionize notebooks into hardened microservices.
• Security & compliance hardening: Implement automated scanning and guardrails (examples mentioned, text cut off).

Requirements

• Strong end-to-end ownership of production AI/ML lifecycle, especially LLMOps/MLOps in multi-cloud environments.

Nice-to-haves (implied by responsibilities)

• Experience deploying RAG systems with vector databases and semantic indexing.
• Experience with Terraform/CloudFormation, observability, and evaluation frameworks for LLMs.
• Familiarity with cloud security/IAM networking patterns (e.g., VPC peering, firewall rules).
• Experience with inference optimization and container/serverless serving.
• Experience integrating enterprise analytics platforms (e.g., Snowflake/Databricks/Palantir).