Senior AWS Generative AI Solution Engineer

Leidos is seeking an experienced Senior AWS Generative AI Solution Engineer to design and implement the reasoning layer for a GenAI-enabled HUD FHA Resource Center. The selected candidate will lead development of a multi-agent solution using Amazon Bedrock, AgentCore Runtime, Knowledge Bases, Guardrails, OpenSearch Serverless, AWS Lambda, and Retrieval-Augmented Generation to support grounded voice and chat interactions.

Design and implement Amazon Bedrock Agents, including a Supervisor Agent and domain-specific Sub-Agents. Develop Retrieval-Augmented Generation workflows using Amazon Bedrock Knowledge Bases and OpenSearch Serverless. Implement prompt engineering, agent routing, confidence logic, and low-confidence fallback behavior. Configure Bedrock Guardrails for PII redaction, prompt-injection detection, denied topics, content safety, and grounding evaluation. Build Lambda orchestration to connect Lex V2, Bedrock AgentCore, response generation, and human transfer decisions. Support session management, conversation context persistence, and cross-channel consistency for voice and chat. Integrate the GenAI stack with Amazon Connect, Salesforce, CloudWatch, DynamoDB, S3, SNS, IAM, and KMS. Lead AI testing, response quality tuning, grounding validation, and latency optimization across production-like scenarios. Please Note: This is 100% telework/remote opportunity.  However, candidate must be/reside in the Gaithersburg, MD/Washington, DC metro area.

Bachelor’s degree and 8+ years of relevant experience OR Masters degree with 6+ years of relevant experience. Additional years of relevant experience may be considered and/or accepted in lieu of degree. 8+ years of software engineering OR cloud development experience.  At least 3 years of Generative AI OR LLM application development experience. Strong Python development skills and experience building production AWS solutions. Hands-on experience with Amazon Bedrock, RAG patterns, and vector-based retrieval. Experience with APIs, distributed systems, and secure cloud architecture. Ability to work effectively in a regulated government environment. Ability to obtain and maintain a Public Trust security clearance. US Citizen.

AWS certifications such as AI Practitioner, Machine Learning Specialty, OR Solutions Architect Professional. Experience with Amazon Bedrock Agents, AgentCore Runtime, Guardrails, and Knowledge Bases. Experience with Claude models, agent frameworks, or multi-agent orchestration patterns. Familiarity with FedRAMP environments and government AI governance requirements. Experience with LangChain, Strands Agents, or other agent development frameworks. Success Measures: Supervisor Agent and Sub-Agents are implemented and tested across HUD service domains. Responses are grounded, accurate, and safely filtered through Guardrails. Confidence scoring and fallback behavior support reliable human escalation. The GenAI platform is production-ready, maintainable, and compliant with security and privacy requirements. Target Technical Skills:  Highly Desired/Preferred Skills (Nice to Have): Amazon Bedrock, Bedrock Agents, Bedrock Guardrails, Bedrock Knowledge Bases, AgentCore Runtime. Retrieval-Augmented Generation, prompt engineering, multi-agent systems, vector search. AWS Lambda, Amazon Lex V2, Amazon Connect, CloudWatch. Python, REST APIs, JSON, CloudFormation or Terraform, CI/CD. OpenSearch Serverless, DynamoDB, S3, IAM, KMS. Please Note:  The program budget salary for this role could fall anywhere between $130,000 to low/mid $150,000 with a slight wiggle room (no guarantees) based on relevant experience and assessment.  Additional factors considered in extending an offer include (but are not limited to) responsibilities of the job, education, experience, knowledge, skills, and abilities, as well as internal equity, alignment with market data, applicable bargaining agreement (if