Indian Institute of Technology Delhi

Executive Certificate in Agentic AI Systems: Foundation to Deployment (AAIS Batch 1)

Indian Institute of Technology Delhi

eVIDYA

Application Deadline Approaching

Last date to apply: January 8, 2027 · Program starts on January 9, 2027

About Programme

The Executive Certificate in Agentic AI Systems (Foundation to Deployment) from CEP, IIT Delhi is designed to equip professionals with the knowledge and practical skills required to design, develop, and deploy autonomous AI systems for enterprise applications. Moving beyond conventional Generative AI, the programme offers a comprehensive learning journey across large language models (LLMs), autonomous AI agents, multi-agent orchestration, Retrieval-Augmented Generation (RAG), AI TRiSM (Trust, Risk & Security Management), and production-ready deployment frameworks.

Through live faculty-led sessions, hands-on learning, and an industry-relevant capstone project, participants gain the expertise to build intelligent, secure, and scalable AI solutions that address real-world business challenges while enabling responsible enterprise AI adoption.

Programme Content

1. Pre-read Module

Python for Data Science | Async pre-read content

Exploratory Data Analysis and Data Visualization | Async pre-read content

Machine Learning | Pre-read content

Deep Learning | Async pre-read content

2. Agentic AI Ecosystem & Foundational Landscape

Evolution of AI: From GenAI to Agentic AI

Building the Agentic AI Mindset & Opportunities

Introduction to LangChain & LangGraph for Agentic Design

Demonstration | Learning Objective | Required Tools

Demo 1: Your First AI Assistant

Learning Objective: Build a simple conversational assistant with a system prompt and configurable model settings.

Required Tools: LLM APIs, Streamlit

Demo 2: Prompt Playground

Learning Objective: Explore how prompts, temperature, and context influence model behavior.

Required Tools: Python, OpenAI

Demo 3: Token & Context Explorer

Learning Objective: Visualize tokenization and context window limitations.

Required Tools: Tiktoken

3. Building with LLMs: Prompt Engineering & Intelligent AI Systems

LLM Foundations (Transformers, Tokens, Context Windows & Scaling Laws)

Prompt Engineering (Chain-of-Thought, Few-Shot & Structured Output)

LangChain LCEL (Chains, Output Parsers & Callbacks)

System Instructions, Prompt Security & Injection Defence

Demonstration | Learning Objective | Required Tools

Demo 1: Email Writing Assistant

Learning Objective: Generate professional emails from short instructions.

Required Tools: LangChain

Demo 2: Resume Improvement Assistant

Learning Objective: Compare a resume with a job description and recommend improvements.

Required Tools: LangChain

Demo 3: Structured Data Extraction

Learning Objective: Extract structured JSON from invoices or forms.

Required Tools: LangChain

Demo 4: Prompt Injection Defense

Learning Objective: Demonstrate common attacks and practical mitigation strategies.

Required Tools: Guardrails

4. Designing & Engineering Autonomous AI Agents

Agent Lifecycle: Observe – Plan – Act – Reflect Loop & State Management

Tool Use, Function Calling & Model Context Protocol (MCP)

Planning, Execution & Memory Systems

Demonstration | Learning Objective | Required Tools

Demo 1: Tool-Calling Weather Agent

Learning Objective: Answer weather queries by invoking external tools.

Required Tools: Function Calling

Demo 2: Personal Task Agent

Learning Objective: Maintain tasks with memory and planning.

Required Tools: Memory, MCP

Demo 3: Travel Planning Agent

Learning Objective: Create travel itineraries using external services.

Required Tools: LangGraph

5. Retrieval-Augmented Generation, Memory & Knowledge Systems

Vector Databases, Embeddings & Similarity Search in Production

RAG in Action: Enterprise Search & AI Knowledge Systems

GraphRAG & Multi-Source Knowledge (SQL, APIs & Real-Time Feeds)

Demonstration | Learning Objective | Required Tools

Demo 1: PDF Question Answering

Learning Objective: Answer questions from uploaded documents with citations.

Required Tools: FAISS/Chroma DB

Demo 2: Enterprise Knowledge Assistant

Learning Objective: Build an internal knowledge chatbot.

Required Tools: RAG

Demo 3: SQL Analytics Agent

Learning Objective: Convert natural language questions into SQL queries.

Required Tools: LangChain

6. Multi-Agent AI Systems & Workflow Orchestration

LangGraph: StateGraph, Nodes, Conditional Routing & Checkpointing

Multi-Agent Orchestration with LangGraph & CrewAI

Collaborative AI Agents: Handoffs, Communication & Recovery

Adaptive AI Agents: Planning, Reflection & Autonomous Reasoning

Demonstration | Learning Objective | Required Tools

Demo 1: Research Team

Learning Objective: Searcher, summarizer, and report writer collaborate.

Required Tools: CrewAI

Demo 2: Software Engineering Team

Learning Objective: Planner, coder, tester, and reviewer solve a coding task.

Required Tools: LangGraph

Demo 3: Customer Support Workflow

Learning Objective: Specialized agents coordinate issue resolution.

Required Tools: CrewAI

7. AI TRiSM: Trust, Risk & Security for Autonomous AI

AI TRiSM Framework for Agentic & Autonomous Systems

Components of AI TRiSM (Model Risk, Hallucination Quantification & Calibration)

Agentic Security (Data Poisoning & Model Inversion)

Responsible Deployment, Explain ability & Guardrail Architecture

Demonstration | Learning Objective | Required Tools

Demo 1: Prompt Injection Detector

Learning Objective: Identify and block malicious prompts.

Required Tools: Guardrails

Demo 2: Hallucination Evaluation

Learning Objective: Assess factual grounding of model responses.

Required Tools: DeepEval

Demo 3: PII Protection

Learning Objective: Detect and mask sensitive information before model invocation.

Required Tools: Presidio

8. From Prototype to Production: Building & Deploying Agent APIs

Agent APIs with FastAPI & Enterprise Service Exposure

Lang Smith: Tracing, Debugging Multi-Step Failures & Agent Run Replay

Async API Design, Streaming Responses & WebSocket Support

Demonstration | Learning Objective | Required Tools

Demo 1: FastAPI Agent Service

Learning Objective: Expose an AI agent through REST APIs.

Required Tools: FastAPI

Demo 2: Web Interface

Learning Objective: Deploy an interactive assistant with Streamlit.

Required Tools: Streamlit

Demo 3: Docker Deployment

Learning Objective: Package the application for portable deployment.

Required Tools: Docker

9. LLMOps & Enterprise Operations, Monitoring, Scaling & Cost Management

LangGraph Cloud Deployment, Stateful Session Management & Scaling

CI/CD for LLM Apps (Prompt Versioning, A/B Testing & Staged Rollouts)

Agentic AI Economics: Cost Optimization, Token Budgeting & ROI Measurement

Demonstration | Learning Objective | Required Tools

Demo 1: LangSmith Tracing and Debugging

Learning Objective: Trace an LLM application and inspect prompts, responses, tool calls, execution steps, token usage, latency, and errors.

Required Tools: LangSmith

Demo 2: Token Cost and Latency Dashboard

Learning Objective: Collect and visualize token consumption, estimated cost, request volume, errors, and response latency.

Required Tools: Streamlit + LangSmith or LiteLLM telemetry

Demo 3: Multi-Model Routing and Failover

Learning Objective: Route requests according to task complexity, cost, latency, and availability; demonstrate retries and model fallbacks.

Required Tools: LiteLLM

10. Assignments / Case-Projects

Prompt engineering challenge on a business dataset

Build and test a single autonomous AI agent

Implement a RAG pipeline for enterprise document Q&A

Design a multi-agent workflow using LangGraph or CrewAI

Apply context engineering techniques to optimize LLM-based applications

Beginner Projects

AI Email Productivity AssistantBuild an assistant that summarizes emails, drafts context-aware replies, and requests approval before sending.

Meeting Intelligence AssistantConvert meeting transcripts into summaries, action items, and follow-up tasks.

Resume Review AssistantEvaluate resumes against job descriptions and recommend targeted improvements.

Enterprise Document Q&ACreate a retrieval-augmented assistant that answers questions from uploaded documents with citations.

Natural Language SQL AnalystTranslate business questions into SQL queries and visualize the retrieved data.

11. Capstone Project

Build an end-to-end Agentic AI application incorporating AI TRiSM principles applications

Programme Audience

Eligibility Criteria

Any science, engineering, or commerce graduate, At least 1 year of experience is preferred (optional)

Programme Benefits

Comprehensive Agentic AI Mastery

AI TRi SM For Responsible AI

Learn From IIT Delhi Faculty

Master Leading Agentic AI Frameworks

Build Enterprise Knowledge Systems

Build An End-to-End Capstone Project

Optional Campus Immersion

Flexible Six-Month Learning Journey

Other programs in this subject area you might find useful

Same topic, Similar duration - Broader exploration across all institutes

Contact us for the further details

Speak with an Advisor

  1. Pedagogy (case method, experiential learning, coaching, simulations)
  2. Programme Objective
  3. Contact Person Details
  4. Testimonials
  5. Brochure

Prashansa Uttam

Programme Advisor

+91 9403890085[email protected]Mon – Fri, 9am – 5pm IST

Indian Institute of Technology Delhi

Executive Education Office

NA
NA
NA
https://home.iitd.ac.in
New Delhi, India

Tell us about your program enquiry

Fill out the form below and our team will get back to you within 24 hours.

Executive Certificate in Agentic AI Systems: Foundation to Deployment | IIT Delhi