Artificial Intelligence Architect
2 days ago
City of London
AI Architect (Microsoft Tech Stack) — Contract Location: London (Hybrid) Contract: 6–12 months (extendable) Start: ASAP IR35: TBD Day Rate: Competitive / Market rate About the Role We’re seeking an experienced AI Architect to lead the design, delivery, and governance of AI solutions across enterprise workloads using the Microsoft ecosystem. You’ll define AI architecture patterns, oversee solution design, guide engineering teams, and ensure alignment with security, compliance, and responsible AI standards. The role spans GenAI, ML, data platforms, MLOps, prompt engineering, and application integration. Key Responsibilities • Architecture & Design, • Define end-to-end AI reference architectures across data, model lifecycle, APIs, and app integrations using Azure services (Azure OpenAI, Azure ML, Azure Databricks, Azure Data Lake, Synapse/Fabric, AKS/Container Apps, Event Hub/Service Bus)., • Select and benchmark models (OpenAI, Azure AI Foundry, OSS LLMs) for task fit, cost, latency, and accuracy; design RAG and fine‑tuning strategies., • Establish MLOps / LLMOps pipelines (CI/CD, feature stores, monitoring, drift detection, human-in-the-loop)., • Delivery Leadership, • Translate business objectives into solution roadmaps, epics, and architecture runway; oversee PoCs → pilots → production., • Provide hands-on guidance to data engineers, ML engineers, and app teams; unblock delivery and manage technical risks., • Data & Integration, • Partner with data teams to design secure data ingestion, curation, and governance in Azure Data Lake/Fabric., • Architect API layers (REST/GraphQL), microservices, and event-driven patterns for AI-enabled apps (including Microsoft 365/Graph integrations where relevant)., • Security, Compliance & Responsible AI, • Implement Azure security best practices: identity (Entra ID), network isolation, secrets (Key Vault), encryption, logging (Azure Monitor)., • Ensure regulatory compliance (GDPR) and responsible AI guardrails (safety filters, content moderation, auditability, model cards)., • Operational Excellence, • Define SLOs/SLAs for AI services; capacity planning, cost management (FinOps), observability, fallback strategies., • Drive governance: model registry, prompt libraries, evaluation harnesses, data lineage, policy enforcement., • Stakeholder Engagement, • Work closely with product, risk, and business sponsors; present architecture decisions and trade-offs to senior stakeholders., • 10+ years in software/data/ML engineering with 5+ years in solution/enterprise architecture, focusing on AI/ML., • Expert with Microsoft Azure AI stack:, • Azure OpenAI Service, Azure AI Foundry, Azure Machine Learning, Azure Databricks, Azure Data Lake Storage, Azure Synapse/Microsoft Fabric., • Azure Kubernetes Service/Container Apps, Key Vault, Event Hub, Service Bus, API Management, App Service, Functions., • Proven delivery of GenAI solutions (RAG, prompt engineering, tools/agents, evaluation frameworks) and ML systems (training, deployment, monitoring)., • Strong in MLOps/LLMOps: pipelines (GitHub Actions/Azure DevOps), model registry, feature store, automated evaluation, canary/blue‑green deploys., • Solid grasp of data architecture: Lakehouse, delta tables, medallion patterns, governance (Purview), performance tuning., • Practical programming proficiency: Python (ML pipelines, SDKs), SQL, and familiarity with .NET/C# or Node/TypeScript for service integration., • Deep understanding of security, compliance, and responsible AI principles; experience in regulated environments is a plus., • Experience with Microsoft 365 Copilot extensibility, Graph connectors, and Teams integrations., • Knowledge of vector databases (Azure AI Search, Redis, Cosmos DB, or pgvector) and embedding strategies., • Exposure to OSS LLMs (Llama, Mistral) and hybrid architectures combining proprietary and open models., • Industry certifications: Azure Solutions Architect Expert, Azure Data Engineer/AI Engineer, DP‑900/AI‑900., • FinServ, Public Sector, or Healthcare domain expertise., • Background in A/B testing, offline & online evaluation (hallucination, faithfulness, toxicity), and cost/performance optimization.