Aurora
Job Description AI Architect Atlanta, GA Job Description Roles & Responsibilities Key Responsibilities Enterprise AI Architecture • Define end-to-end architecture for Generative AI, Agentic AI, machine learning, and intelligent automation solutions., • Translate airline business priorities into scalable AI capabilities, reference architectures, solution patterns, and implementation roadmaps., • Design reusable AI services across digital channels, airline operations, IT operations, customer service, engineering, and enterprise functions., • Establish architecture standards for model integration, orchestration, data access, APIs, security, observability, evaluation, and deployment., • Review solution designs and ensure alignment with enterprise standards and target architecture. Generative AI and Agentic AI • Architect enterprise-grade LLM solutions using RAG, knowledge grounding, tool integration, and multi-agent orchestration., • Design autonomous and human-in-the-loop workflows with clear controls, approvals, escalations, and auditability., • Define patterns for agent planning, reasoning, state management, memory, function calling, structured outputs, and secure tool access., • Evaluate frameworks such as LangGraph, LangChain, Semantic Kernel, CrewAI, or equivalent enterprise technologies., • Design reusable components for prompts, tools, workflows, model gateways, evaluation, guardrails, and agent observability. Airline Business and Operational Solutions • Partner with airline business and technology teams to identify and prioritize high-value AI opportunities., • Architect solutions for digital customer experience, personalization, operational reliability, disruption management, reservations, customer service, employee assistance, major incident management, intelligent IT operations, engineering productivity, and knowledge management., • Design AI capabilities for high-volume, near-real-time, customer-facing, and operationally critical environments., • Balance innovation and speed with availability, reliability, safety, security, and operational stability. Data, Context, and Knowledge Architecture • Design secure data and knowledge architectures that ground AI solutions in trusted enterprise information., • Define patterns for ingestion, chunking, metadata, embeddings, vector search, reranking, retrieval, response validation, and knowledge freshness., • Integrate structured, unstructured, streaming, and operational data through enterprise data platforms, APIs, and event streams., • Partner with data teams to ensure quality, lineage, access control, privacy, and governance for information used by AI systems. AWS Cloud and AI Platforms • Architect AI solutions using Amazon Bedrock, SageMaker, OpenSearch, S3, EKS/ECS, Lambda, Step Functions, API Gateway, Glue, Athena, Redshift, Kinesis/MSK, IAM, KMS, Secrets Manager, and CloudWatch., • Evaluate models and services based on quality, security, latency, scalability, portability, reliability, and cost., • Design cloud-native AI platforms that support experimentation, controlled de ployment, enterprise reuse, and model choice., • Integrate third-party and open-source models where appropriate while maintaining enterprise security and governance. AI Engineering and Integration • Provide hands-on architecture guidance for Python-based AI services, APIs, microservices, and event-driven applications., • Define secure integration patterns for agents to interact with enterprise applications, APIs, databases, and operational tools., • Establish standards for schema validation, exception handling, retries, fallbacks, rate limits, and human escalation., • Guide teams in building modular, testable, reusable, and maintainable AI components. LLMOps, MLOps, Observability, and Production Readiness • Define lifecycle standards for model selection, prompt management, training, fine-tuning, testing, deployment, monitoring, versioning, and retirement., • Establish CI/CD and automated testing for models, prompts, retrieval pipelines, agents, APIs, and supporting services., • Design evaluation frameworks covering accuracy, groundedness, relevance, safety, latency, reliability, operational impact, and cost., • Implement end-to-end tracing and observability for model calls, retrieval decisions, agent execution, tool usage, and failures., • Define production-readiness criteria, rollback approaches, service objectives, support models, and incident-response procedures. Responsible AI, Security, and Governance • Embed responsible AI, privacy, cybersecurity, compliance, and risk controls into architecture and delivery., • Define identity, authorization, encryption, data isolation, auditability, secrets management, and sensitive-data protection controls., • Implement guardrails for prompt injection, hallucination, data leakage, unsafe output, and unauthorized tool execution., • Ensure traceability and appropriate human oversight for high-impact or operationally sensitive AI recommendations and actions. Technical Leadership and Collaboration • Serve as a trusted AI architecture advisor to business and technology leadership., • Lead architecture reviews, technical workshops, design sessions, and AI use-case assessments., • Provide clear recommendations on technology selection, implementation approach, risks, dependencies, and trade-offs., • Mentor AI engineers, data scientists, software engineers, and solution architects., • Create reference architectures, reusable patterns, standards, playbooks, and contributions to the enterprise AI roadmap.