AI Infrastructure Engineer

Role: AI Infrastructure Engineer Location: Remote Duration: 12+ Months Contract Visa: US Citizen Job Summary: Defense Counterintelligence and Security Agency (DCSA) has responsibility for operations and maintenance of nine applications supporting the Background Investigation Enterprise Systems (BIES). The program supports modernization efforts by migrating systems to a secure cloud environment and modernizing software that captures, processes, and retains background investigation information for personnel seeking access to sensitive information Task Description: Define IAM roles, VPC endpoints, Knowledge Bases, and Guardrails as Terraform modules - not console clicks.Version and distribute reusable infrastructure artifacts (Terraform modules, container images, Lambda packages) through Artifactory.Build and deploy connector infrastructure (e.g., MCP servers, Lambda-based integrations) linking Bedrock to internal APIs, databases, and ticketing systems — hosted and version-controlled like any other service.Deploy custom/fine-tuned models via Bedrock Custom Model Import or Provisioned Throughput, including the cost/commitment tradeoff analysis that comes with it.Build cost/usage observability (per-model, per-developer spend and budget alerts) so spend doesn't scale silently Required skills/Level of Experience: Active Secret clearance with Security Plus certification. U.S. Citizenship required.Terraform - Bedrock resources (IAM, Knowledge Bases, Guardrails, VPC endpoints, Provisioned Throughput) as reusable, version-controlled modules.Artifactory - manages and distributes versioned infrastructure artifacts for reproducibility across dev/test/prod.Custom/fine-tuned model deployment - hands-on with Bedrock Custom Model Import (open-source model weights, size/format constraints) and Provisioned Throughput commitment planning.Connector/integration hosting - builds and operates the actual services (Lambda, ECS, containers) exposing internal data sources to Bedrock agents - not just configuring existing ones see additional comments.Experience-2+ years Nice to have skills: Reinforcement fine-tuning (RFT) workflow experience on Bedrock.Application Inference Profiles for per-team cost attribution.Multi-account AWS Organizations structure for dev/test/prod separation.CI/CD pipeline integration for Terraform + Artifactory promotion workflows.Cross-cloud IaC experience (Azure/GCP) supporting anti-lock-in posture.