AI Infrastructure Architect

About Accenture

Qualifications

• Bachelor's degree or equivalent in Computer Science, Computer Engineering, Data Science, AI/ML, Information Technology or a related engineering discipline.
• Minimum 15+ years of overall experience across software engineering, data engineering, AI/ML engineering, cloud architecture, enterprise architecture or technology leadership.
• Minimum 8+ years of experience designing and deploying enterprise-grade advanced AI, data, analytics or cloud-native solutions using at least one cloud vendor.
• Minimum 2+ years of experience in LLM and generative AI solution architecture, including agentic systems, RAG, prompt engineering, model integration and evaluation patterns.
• Minimum 2+ years of experience architecting and operationalizing LLM-driven application architecture patterns in enterprise-scale or production environments.
• Minimum 6+ years of experience in engineering, machine learning, deep learning, NLP solutions, data engineering or large-scale analytical engineering applications.
• Minimum 6+ years of experience as a machine learning / data / AI architect designing large-scale analytical engineering solutions in industry contexts such as banking, insurance, healthcare, retail, telecom or capital markets.
• Required Skills/ Experience
• Deep architecture and hands-on engineering experience with Amazon Bedrock, Bedrock Agents/AgentCore, Knowledge Bases, Bedrock Guardrails, Bedrock model evaluation, SageMaker, Lambda, API Gateway, Step Functions, EventBridge, OpenSearch Serverless/Vector Engine, S3, IAM, VPC, KMS, CloudWatch, CloudTrail and AWS Well-Architected practices.
• Strong expertise in enterprise AI platform architecture covering RAG, embeddings, vector databases, semantic retrieval, context engineering, model routing, agent orchestration, memory, tool calling, AI gateways and model evaluation.
• Ability to set enterprise NFRs and architectural controls for performance, scalability, security, privacy, reliability, governance, observability, resiliency, cost optimization and operational readiness.
• Experience making definitive, evidence-based decisions on design patterns, reference architectures, frameworks, technology selections, foundation models and deployment approaches.
• Experience establishing agent registry and certification models, AI control planes, access models, guardrails, production evaluation stacks, model risk controls and cross-platform governance.
• Strong executive communication, architecture governance and thought leadership skills with ability to influence business, technology, security, product and delivery leadership teams.
• Good to Have Skills
• AWS Certified Solutions Architect Professional, Machine Learning Specialty or Generative AI related certification experience with AWS CDK/Terraform, EKS, Bedrock AgentCore, Amazon Q, private connectivity, landing zones, regulated workloads and enterprise FinOps.
• Exposure to open-source AI and orchestration frameworks such as LangChain, LangGraph, LlamaIndex, Haystack, Semantic Kernel, MLflow, FastAPI, Docker and Kubernetes.
• Experience with responsible AI, model risk management, AI governance boards, red-teaming, synthetic data, human-in-the-loop review, A/B testing and GenAI FinOps.
• Recognized thought leadership through enterprise reference architectures, internal capability building, client advisory, platform accelerators, publications, whitepapers, conference sessions or industry forums.
• 15 years full time education

Responsibilities

• Partner with client executives and business leaders to define the enterprise AI strategy, target-state architecture and investment roadmap across platforms, data, models, applications and operating model.
• Lead enterprise AI assessments, technology comparisons, platform selection, reference architecture definition, modernization opportunities and implementation sequencing for complex transformations.
• For AWS, set the AWS-native enterprise AI platform strategy define Bedrock-led foundation architecture for GenAI applications, federated agent systems, AI gateway patterns, governed RAG and secure model inference make build-versus-buy decisions across Bedrock, SageMaker and open-source frameworks establish multi-account, security, observability, resilience and GenAI FinOps standards.
• Own the complete end-to-end technical solution for complex AI platforms, ensuring every domain is designed cohesively against business objectives, enterprise standards and non-functional requirements.
• Translate governing architecture principles into a concrete, defensible technical solution that platform, data, AI/ML and application engineering teams can build against.
• Set architectural direction for model- and tool-agnostic multi-agent systems, including orchestration, memory, tool/skill use, agent registry, AI gateway/control plane, risk scoring and certification gates.
• Define the enterprise context layer architecture across knowledge graphs, ontologies, vector search, semantic retrieval, prompt/context assembly, conversational state and reusable memory services.
• Establish identity, authorization, layered guardrails, prompt-injection defense, PII protection, audit logging, lineage and defense-in-depth controls for AI agents, tools, data and models.
• Mandate productized evaluation and observability practices covering accuracy, relevance, groundedness, model quality, latency, cost, safety, reliability, production support and continuous improvement.
• Establish FinOps as a first-class AI concern, including usage labelling, token budgets, gateway-enforced budgets, cost-per-archetype planning, threshold alerts and optimization levers.
• Produce and steward authoritative architecture assets including enterprise AI blueprints, ADRs, sequence diagrams, solution patterns, interface specifications, reference architectures and governance playbooks.