DSCA AI Engineer - Generative AI, RAG & Agentic Systems (Software Development Analyst) - 30714

Job Description

We are seeking an experienced AI/RAG Engineer to design, build, and optimize secure generative AI, retrieval-augmented generation (RAG), agentic, and AI-governance capabilities for an enterprise application in an AWS GovCloud IL4/IL5 environment. 

The engineer will lead development of the AI intelligence layer, including data ingestion, retrieval, model integration, prompting, agent workflows, evaluation, AI gateway configuration, and AI security. The role will work closely with front-end, back-end, platform, cybersecurity, and Department of War (DoW) identity teams to deliver reliable production AI capabilities. 

Essential Job Responsibilities

• Design, develop, and optimize production generative AI, RAG, and agentic capabilities using approved Amazon Bedrock services. 

• Evaluate and integrate foundation models based on quality, relevance, latency, reliability, security, and cost. 

• Build and improve RAG pipelines supporting ingestion, preprocessing, chunking, embeddings, vector search, retrieval, reranking, and context generation. 

• Design agentic workflows that combine foundation models, retrieved knowledge, approved tools, and enterprise APIs to complete bounded multi-step tasks. 

• Develop prompts, structured-output patterns, tool or function-calling interfaces, and guardrail configurations appropriate to application requirements. 

• Establish AI evaluation methods, test datasets, and monitoring approaches to assess response quality, groundedness, accuracy, safety, tool-use outcomes, latency, reliability, and cost. 

• Develop knowledge-ingestion pipelines for structured and unstructured data while preserving source-level access controls and applicable data-handling requirements. 

• Implement AI security controls, including prompt-injection defenses, input and output validation, data protection, audit logging, and human-review or escalation workflows where required. 

• Design and administer AI gateway configurations for approved model and agent access, including model routing and allowlists, service-identity or virtual-key controls, rate limits, budgets, usage attribution, fallback behavior, logging, and guardrail integration. 

• Develop Python-based AI services, APIs, and integration patterns for consumption by backend application services. 

• Partner with platform teams on infrastructure-as-code, deployment, observability, and operational support for AI/RAG services, AI gateways, ingestion pipelines, vector-search capabilities, and supporting AWS resources. 

• Collaborate with backend, front-end, cybersecurity, platform, and identity teams to integrate secure and dependable AI capabilities into production applications. 

Minimum Qualifications

• Strong Python development experience. 

• Experience building and operating production generative AI, RAG, or AI-enabled applications. 

• Strong understanding of large language models, embeddings, vector search, retrieval patterns, prompt design, structured outputs, and AI service APIs. 

• Experience designing RAG pipelines and integrating models with enterprise data sources, tools, APIs, or application services. 

• Experience developing or evaluating agentic workflows that use models, retrieval, and approved tools to complete bounded tasks. 

• Hands-on experience with Amazon Bedrock, Boto3, and AWS cloud services. 

• Experience evaluating and improving AI capabilities for quality, groundedness, latency, reliability, security, and cost. 

• Experience implementing practical AI security and governance controls, including prompt-injection mitigation, input and output validation, access-control integration, rate limiting, usage monitoring, logging, and data protection. 

• Experience with Git-based development, automated testing, and collaborative software-delivery practices. 

• Clearance: Must possess and maintain an active Secret clearance at the time of consideration
• Fully Remote position; must align with Eastern Time Zone core business hours

***Salary level based on background, direct experience and education

Preferred Requirements

• 8+ years of preferred experience 
• Experience with Amazon Bedrock AgentCore or comparable agent-runtime, orchestration, memory, tool-connectivity, or observability platforms. 

• Experience with LiteLLM or comparable AI gateway, model-proxy, or AI-governance platforms supporting multi-model routing, access controls, budgets, observability, and auditability. 

• Experience working in AWS GovCloud or other federal, defense, intelligence, or regulated cloud environments. 

• Experience with OpenSearch, pgvector, or comparable vector-search and retrieval platforms. 

• Familiarity with AWS CDK, CloudFormation, Terraform, or equivalent infrastructure-as-code tools. 

• Familiarity with DoW identity and access-management concepts, Zero Trust principles, least-privilege IAM, and secure service-to-service communication. 

• Familiarity with DISA STIGs, SRGs, RMF activities, security-control evidence, continuous monitoring, or vulnerability-remediation workflows. 

• Experience with LangChain, LlamaIndex, MCP-compatible tool integrations, or comparable AI orchestration frameworks. 

• Experience working with controlled, sensitive, or otherwise regulated data in a federal, defense, intelligence, or similarly regulated environment. 

Physical Requirements

• Ability to remain in a stationary position for prolonged periods while working at a computer.
• Ability to operate a keyboard, mouse, and standard office equipment to perform analytical tasks.
• Must be able to effectively communicate in virtual meetings, briefings, and secure collaboration environments.
• Ability to review, analyze, and interpret digital data for extended periods with attention to detail and accuracy.
• Occasional light lifting of documents or equipment up to 10 lbs, if needed for home-office setup.
• Must maintain a home work environment that meets company and government security requirements (e.g., controlled space for handling sensitive or FOUO information, if applicable).
• Must be able to work assigned East Coast core business hours to support program and customer requirements.
• May require <10% travel to customer meetings (East Coast).