New Delhi: As artificial intelligence advances from generating content to executing increasingly complex tasks, enterprises face a new set of challenges in developing systems that can operate with greater autonomy while maintaining reliability and oversight. Against this backdrop, the Continuing Education Programme (CEP), Indian Institute of Technology Delhi (IIT Delhi), has announced the inaugural batch of its Executive Certificate in Agentic AI Systems (Foundation to Deployment). The programme seeks to develop the technical capabilities professionals need to design, engineer and deploy autonomous AI solutions within enterprise environments. The six-month programme examines the architecture and practical application of agentic AI, extending beyond conventional generative AI to explore how intelligent systems reason, plan, access information and coordinate actions. Its curriculum encompasses large language models (LLMs), autonomous AI agents, multi-agent orchestration, Retrieval-Augmented Generation (RAG) and production deployment. Participants will work with frameworks and protocols such as Model Context Protocol (MCP), LangGraph, CrewAI and AutoGen. A dedicated focus on AI Trust, Risk and Security Management (AI TRiSM) will address model governance, explainability, security, compliance and the safeguards organisations must consider when deploying autonomous systems. Speaking on the announcement of the inaugural batch, Professor Sandeep Kumar, Associate Professor – Department of Electrical Engineering, School of Artificial Intelligence, IIT Delhi, said, “The challenge is no longer simply to make AI generate better answers. It is to build systems that can take action, use information responsibly and remain dependable when conditions change. We want participants to understand those choices, test their assumptions and develop applications that organisations can confidently use in practice.” Industry projections indicate that autonomous AI systems could assume a considerably larger role in enterprise operations. Gartner estimates that agentic AI will resolve 80 per cent of common customer service issues without human intervention by 2029. It also forecasts that 40 per cent of enterprise applications will incorporate task-specific AI agents by the end of 2026. The programme caters to AI and data professionals, technology leaders, software developers, IT practitioners and science and engineering graduates seeking specialised expertise in agentic AI. Eight core modules cover the foundations of agentic AI, prompt engineering and intelligent systems, autonomous agent design, enterprise knowledge systems, multi-agent orchestration, AI TRiSM, application deployment and enterprise-scale operations. Participants will undertake assignments and projects of increasing complexity, progressing from individual AI assistants and knowledge-retrieval applications to multi-agent workflows. An end-to-end capstone project will require them to apply their learning to an agentic AI application while incorporating AI TRiSM principles. Throughout the programme, learners will gain experience with relevant development frameworks, libraries and deployment tools.
