Expert-level experience in AI/ML solution architecture in enterprise environments
Strong proficiency in Azure Cloud, preferably Azure AI Foundry (or equivalent: AWS Bedrock, Vertex AI)
Deep knowledge of LLMs (OpenAI, Anthropic Claude, Mistral, etc.) and their operational characteristics
Proven experience with RAG and Agentic RAG (chunking, embeddings, vector stores such as Azure AI Search, pgvector)
Hands-on experience with orchestration frameworks: LangChain, LangGraph, Semantic Kernel, Microsoft Agent Framework
Strong programming skills in Python; solid understanding of SOLID principles and design patterns for AI systems
Experience with Event-Driven Architecture and API integrations (REST/GraphQL)
Knowledge of Kubernetes and DevOps/MLOps practices (Docker, Terraform, GitHub Actions, Azure DevOps)
Strong understanding of security, governance, and compliance in AI systems
Define and own the end-to-end agentic architecture (Agent Platform, Data & Knowledge, Security & Compliance, Governance, Human Oversight, Enterprise Integration)
Design internal SDKs and API contracts ensuring modularity, scalability, and extensibility
Select and justify technology stack (LLM gateway, orchestrator, vector store, cloud platform)
Define integration strategies with enterprise systems (Jira, Confluence, GitHub, Teams, CI/CD pipelines)
Evaluate multi-agent vs. single-agent approaches and document architecture decisions (ADRs)
Ensure compliance by design (EU AI Act) and implement security controls (guardrails, RBAC, data masking, audit trails, prompt injection protection)
Define identity management (Azure AD / OAuth2), agent sandboxing, and human-in-the-loop policies
Lead red teaming exercises and establish risk and IP governance frameworks
Implement quality gates across the agentic SDLC and define success metrics (ROI, cycle time, automation coverage, compliance SLAs)
Oversee observability frameworks including dashboards, logging, and productivity reporting
Mentor and guide GenAI Platform Engineers and Agentic Engineers
Collaborate with Product Owners and stakeholders to prioritize technical backlogs
Act as the senior technical authority in key client decision-making moments
