Azure AI Architect

Tata Consultancy Services; India; On-site; INR 6,000,000 - 9,000,000; Full time

Job summary

Tata Consultancy Services in India seeks an experienced AI architect to define targetstate Azure AI reference architectures using Azure AI Foundry, Azure OpenAI, and Azure AI Document Intelligence for intelligent document processing, enterprise copilots, and knowledge retrieval workflows.

You will lead full-stack AI solution design, with integration patterns, API contracts, data flow, security boundaries, and observability, and establish ADRs, guardrails, and reusable blueprints to accelerate

Responsibilities

• Define targetstate Azure AI reference architectures using Azure AI Foundry, Azure OpenAI, and Azure AI Document Intelligence for intelligent document processing, enterprise copilots, and knowledgeretrieval applications.
• Establish standards for RAG architectures: chunking, embeddings, vector indexing, hybrid ranking, grounding, and citation-based policies using Azure Cognitive Search or vector-enabled data stores.
• Create architectural decision records (ADRs), enterprise guardrails, and reusable blueprints to accelerate solution delivery.
• Lead full-stack AI solution design including integration patterns (Azure Functions, APIM, Logic Apps, Event Grid), API contracts, data flow, security boundaries, and observability.
• Architect document intelligence workflows for classification, extraction, OCR, validation, and seamless integration with CRM/ERP/ITSM/EDM platforms.
• Define and ensure adherence to non-functional requirements availability, latency, scalability, throughput, DR/RTO/RPO, and cost.
• Implement enterprise-grade security: Entra ID, RBAC, network isolation (Private Endpoints), Key Vault secrets, data encryption, governance, and content filtering.
• Embed Responsible AI practices including safety filters, content controls, prompt and data governance, redteaming, and humanintheloop flows.
• Ensure compliance with regulatory standards (GDPR, ISO 27001, legal, risk, and security assessments).
• Define MLOps and LLMOps processes for model versioning, evaluation, deployment, rollback, drift monitoring, and continuous improvement.
• Instrument observability across AI workloads (logging, tracing, metrics, App Insights) and define SLIs/SLOs for mission-critical AI services.
• Drive FinOps discipline—token optimization, model selection, caching, throughput control, and cost monitoring.
• Mentor developers and engineers on prompt engineering, grounding techniques, tool-use patterns, evaluation frameworks, and best practices for LLM design.

Job description

Role & responsibilities

• Define targetstate Azure AI reference architectures using Azure AI Foundry, Azure OpenAI, and Azure AI Document Intelligence for intelligent document processing, enterprise copilots, and knowledgeretrieval applications.
• Establish standards for RAG architectures: chunking, embeddings, vector indexing, hybrid ranking, grounding, and citation-based policies using Azure Cognitive Search or vector-enabled data stores.
• Create architectural decision records (ADRs), enterprise guardrails, and reusable blueprints to accelerate solution delivery.
• Lead full-stack AI solution design including integration patterns (Azure Functions, APIM, Logic Apps, Event Grid), API contracts, data flow, security boundaries, and observability.
• Architect document intelligence workflows for classification, extraction, OCR, validation, and seamless integration with CRM/ERP/ITSM/EDM platforms.
• Define and ensure adherence to non-functional requirementsavailability, latency, scalability, throughput, DR/RTO/RPO, and cost.
• Implement enterprise-grade security: Entra ID, RBAC, network isolation (Private Endpoints), Key Vault secrets, data encryption, governance, and content filtering.
• Embed Responsible AI practices including safety filters, content controls, prompt and data governance, redteaming, and humanintheloop flows.
• Ensure compliance with regulatory standards (GDPR, ISO 27001, legal, risk, and security assessments).
• Define MLOps and LLMOps processes for model versioning, evaluation, deployment, rollback, drift monitoring, and continuous improvement.
• Instrument observability across AI workloads (logging, tracing, metrics, App Insights) and define SLIs/SLOs for mission-critical AI services.
• Drive FinOps discipline—token optimization, model selection, caching, throughput control, and cost monitoring.
• Mentor developers and engineers on prompt engineering, grounding techniques, tool-use patterns, evaluation frameworks, and best practices for LLM design.