Description:

AI is rapidly evolving from simple chat interfaces to autonomous systems capable of planning, reasoning, using tools, and collaborating to achieve complex business objectives. Architecting Autonomy: Agentic AI Bootcamp is an intensive three-week, hands-on program designed for technical professionals who want to move beyond experimentation and build production-ready AI agents within enterprise environments.

Participants will develop a strong foundation in the technologies that power modern agentic systems, including Python, Docker, prompt engineering, LangChain, LangGraph, Model Context Protocol (MCP), and AI-assisted quality engineering. Through a combination of guided instruction, practical labs, and real-world implementation exercises, learners will explore how intelligent agents think, act, interact with enterprise data, and coordinate across complex workflows.

The bootcamp emphasizes architectural thinking, engineering discipline, and operational readiness. Students will learn how to design reliable agent workflows, build retrieval-augmented applications, orchestrate multi-agent systems, expose enterprise capabilities through MCP services, and deploy scalable AI solutions using modern development practices.

Beyond building agents, the program addresses the challenges organizations face when adopting AI on a scale, including testing, governance, security, observability, compliance, and quality assurance. Participants will gain practical experience implementing guardrails, validating AI-generated outputs, and ensuring systems remain trustworthy in production environments.

By the end of the bootcamp, attendees will have designed, built, tested, and deployed a portfolio of agentic AI solutions that demonstrate real business value. Managers will benefit from team members who can confidently evaluate AI opportunities, accelerate innovation, and architect enterprise-ready autonomous systems that align with organizational goals.

By the end of this AI Bootcamp, you will have the confidence and competence to tackle building Agentic AI systems.

Duration: 

3 Weeks

Course Code: BDT636

This bootcamp is designed for individuals with a strong interest in AI and a desire to build AI Agents into existing corporate infrastructure. It might include Technical Leads, Software Engineers, and Solution Architects, QA Engineers.

1. One or more years technical experience
2. Programming experience with Python is must.
3. Basic understanding of using Large Language Models, including prompting

Course Outline:

 Pydantic Library

1. Structured inputs/outputs for tools, agents, and workflows
2. Typed schemas for message passing
3. Models, fields, validators and using BaseModel

AsyncIO Fundamentals

1. Event loops, tasks, coroutines, and concurrency primitives
2. Running parallel tool calls, API requests

Containerization With Docker 

1. Introduction to Containers
2. Docker overview
3. Docker commands
4. Understanding Dockerfile
5. Building Docker Containers
6. Using Docker-Compose for building and testing software

Introduction to Prompt Engineering

Prompt Engineering Foundations:

1. How modern Large Language Models (LLMs) generate responses
2. Tokens, context windows, and model limitations
3. Why prompts matter more than ever in AI-assisted work
4. Common causes of poor responses

The Collaboration Mindset

1. Human-AI collaboration patterns
2. The interactive prompting lifecycle
3. Working with uncertainty and incomplete information
4. Using AI to challenge assumptions
5. Asking for alternatives, trade-offs, and second opinions
6. Understanding confidence versus correctness
7. When human judgment should override AI recommendations
8. Building AI-assisted workflows for daily work

Prompt Refinement Workshop

1. Improve clarity
2. Add business context
3. Add constraints
4. Request alternative solutions
5. Evaluate improvements after each iteration

Core Prompting Techniques

1. Role-Based Prompting
2. Context-Driven Prompting
3. Task-Oriented Prompting
4. Output Structuring
5. Constraint-Based Prompting

Structured Prompt Design

1. Role + Context + Task + Constraints + Output Format

Advanced Prompting Patterns

1. Few Shot Prompting
2. Scenario-Based Prompting
3. Multi-Step Prompting
4. Critique and Revision Prompting
5. Evaluation Prompts
6. Building Reusable Prompt Templates
7. Incident investigation
8. Documentation generation
9. Code review
10. Requirements analysis
11. Executive reporting

Hands-on Labs for topics covered

Practical Agentic AI: Building Intelligent Agents

Foundations of Agentic AI

1. The Agentic Model: think → act → observe → adjust
2. Degrees of Autonomy
3. Agentic AI Applications
4. Decomposing a Task for Agentic AI
5. Common Design Patterns:
6. ReAct
7. Plan–Execute
8. Reflection
9. Overview of the OpenAI SDK: Assistants, messages, and functions
10. The anatomy of an agent loop (reason–act–observe)
11. Practical exposure to tools (retrieval, code execution, function calling)
12. Multiple Hands-on Labs for topics covered

Building Agentic Behavior: Memory, Tools, and Autonomy

1. State and memory: how agents remember context and past actions
2. Tool orchestration: context-aware selection
3. Error handling and guardrails: keeping autonomy reliable
4. Evaluation and logging: traces and workflow analysis
5. Integrating with external data (e.g., retrieval-augmented generation, APIs)
6. Multiple Hands-on Labs for topics covered

Multi-Agents Systems: Coordinating Workflows

1. Planning a practical agent system end-to-end
2. Introduction to LangGraph: graphs as stateful workflows for agents
3. Nodes, edges, and control flow: the logic of orchestration
4. How LangGraph builds on the OpenAI SDK
5. LangGraph for persistent agents, collaboration, workflow management
6. Multiple Hands-on Labs for topics covered

Generative AI Development with LangChain and LangGraph

LangChain Fundamentals

1. What is LangChain?
2. Installing LangChain and setting up environment
3. Creating and using prompts
4. Hands-on: Writing a basic prompt with LangChain

Model Inputs and Outputs

1. Understanding model inputs and outputs
2. Connecting LangChain with an LLM
3. Using prompt templates (zero-shot, one-shot, few-shot learning)
4. Hands-on: Experimenting with model inputs and outputs

Data Connections and Vector Stores

1. Connecting LLMs to external data
2. Using document loaders and transformers
3. Working with text embeddings and vector databases (FAISS, ChromaDB)
4. Hands-on: Implementing a vector database for retrieval

Using Chains for Application Logic

1. Understanding LLM chains
2. Building sequential and router-based chains
3. Hands-on: Developing a LangChain pipeline

Memory in LangChain

1. Introduction to memory and its use cases
2. Types of memory and memory buffers
3. Hands-on: Implementing chat memory for contextual conversations

Building AI Agents with LangChain

1. Introduction to AI Agents and their applications
2. Understanding ReAct framework in LangChain
3. Exploring agent tools and executions
4. Hands-on: Building an agent-based application

Introduction to LangGraph

1. What is LangGraph?
2. Why use LangGraph with AI workflows?
3. Flow engineering and LangGraph components

Building a Reflection Agent

1. What is a Reflection Agent?
2. Creating a reflection chain and defining LangGraph graphs
3. Understanding State Graphs
4. Hands-on: Implementing a Reflection Agent

Developing a Reflexion Agent

1. Understanding Reflexion agent for iterative improvement
2. Build Actor, Revisor agents and integrate message graph
3. Integrating search tools (Tavily, others)
4. Hands-on: Creating a Reflexion Agent with LangGraph

 Advanced RAG (Retrieval-Augmented Generation)

1. Constructing an advanced RAG workflow
2. Using nodes and states for decision making
3. Leveraging predefined LangChain prompt templates
4. Implementing Self-RAG for reducing hallucinations
5. Working with Adaptive RAG for dynamic retrieval
6. Hands-on: Build an advanced RAG system

Persistence & State Management in LangGraph

1. Using state persistence in LangGraph
2. Implementing memory-saving techniques and human-in-loop interventions
3. Storing State using SQLite saver
4. Hands-on: Using persistence in LangGraph application

LangGraph Ecosystem & Cloud Deployment

1. Setting up LangGraph Studio (IDE)
2. LangGraph Cloud API (set up local environment)
3. Using LangGraph Cloud API: Threads, Runs, Assistants
4. Hands-on: Install ecosystem on local machine (Docker required)

Model Context Protocol (MCP) Based AI Systems

MCP and Agentic Architecture Fundamentals

1. Evolution from chatbots to AI agents
2. Why MCP was introduced
3. MCP versus traditional APIs, plugin architecture
4. MCP within modern AI ecosystems
5. Enterprise use cases and architectural benefits
6. Hands-on: Exploring an MCP Ecosystem

MCP Core Components in Depth

1. MCP architecture overview
2. Servers, clients, and transport
3. Tools, Resources, and Prompts
4. Context management concepts
5. Schema-driven communication
6. Transport options (STDIO, SSE, HTTP)
7. Hands-on: Creating your first MCP Server

Structuring Data with Pydantic

1. Schema-first development
2. Pydantic models and validation
3. Input and output contracts
4. Field constraints and metadata
5. Preventing schema drift
6. Hands-on: Building strongly typed tools

Setting Up the MCP Server Environment

1. MCP SDK overview
2. Python project organization
3. Dependency and Configuration management
4. Logging fundamentals
5. Development workflows
6. Hands-on: Building a Maintainable MCP Project

Designing and Implementing MCP Tools

1. Tool design principles and granularity decisions
2. Stateless versus stateful tools
3. Input validation strategies
4. Output enforcement
5. Error handling patterns
6. Tool documentation practices
7. Hands-on: Building Business-Oriented MCP Tools

Building Backend Domain Services

1. Domain-driven design concepts
2. Separating MCP from business logic
3. Service-layer architecture
4. Reusable backend components
5. Hands-on: Creating a Domain Service Layer

Data Access and Persistence

1. In-memory storage patterns
2. SQLite integration, SQL Alchemy ORM
3. Repository pattern
4. Transaction management
5. Async data access, Caching considerations
6. Hands-on: MCP-Powered Data Services

Exposing Enterprise Data Through MCP

1. Tool versus resource design
2. Read-only versus mutating operations
3. Data exposure strategies
4. Access control patterns
5. Working with external REST APIs
6. Hands-on: Building an Enterprise Knowledge Server

MCP and LangChain Integration

1. MCP as a tool provider for agents
2. LangChain tool abstractions
3. Consuming MCP tools from LangChain
4. Structured outputs
5. Memory considerations
6. Hands-on: LangChain Agent Using MCP Tools

Prompts, Context, and Sampling

1. MCP prompt resources and managing prompt templates
2. Dynamic context construction
3. Prompt versioning
4. Improving reliability and consistency
5. Hands-on: Prompt-Driven MCP Applications

Authentication, Authorization, and Security

1. Security considerations for MCP
2. Authentication approaches
3. API keys and tokens
4. OAuth-based integrations
5. Authorization models
6. Secrets management
7. Hands-on: Securing an MCP Server

Testing, Reliability, and Governance

1. Unit testing MCP tools
2. Integration testing
3. Contract testing
4. Failure injection
5. Retry strategies
6. Guardrails and safety controls
7. Auditability considerations

Observability and Monitoring

1. Logging best practices and distributed tracing concepts
2. Metrics collection
3. Tool invocation monitoring
4. Performance analysis
5. Troubleshooting production issues
6. Hands-on: Observing MCP Operations

Packaging, Deployment, and Versioning

1. Packaging with “Pyproject.toml”
2. Dependency management, and semantic versioning
3. Containerization with Docker
4. Deployment strategies
5. Hands-on: Deploying an MCP Service

Multi-Agent Systems and MCP

1. MCP in multi-agent architecture
2. Agent specialization
3. Shared tool ecosystems
4. Coordination patterns
5. Resource sharing
6. Emerging MCP ecosystem trends
7. Hands-on: Multi-Agent MCP Solution

AI-Assisted Testing and Quality Engineering

How AI-Assisted Development Changes Quality Engineering

1. AI-Assisted development shifts delivery patterns and introduces new defect types
2. AI-generated code is often syntactically correct but logically flawed
3. Agent-driven workflows create hidden regression and overconfidence risks
4. Quality roles evolve from defect detection to proactive risk prevention
5. Quality must shift left with earlier guardrails, validation, and oversight

Risk-Based Testing for AI-Generated Code

1. AI delivery risk comes from fast, opaque, unpredictable code changes
2. High-risk AI changes include over-generalized logic, missing boundaries, and hallucinated APIs
3. Requirement-to-risk mapping guides where to focus testing effort
4. Risk driven planning helps prioritize testing under accelerated releases

AI-Assisted Test Design and Automation

1. AI tools accelerate test design and automation across multiple layers
2. AI generates unit, integration, API, UI, contract tests, and documentation
3. Strong prompts and context to improve test generation quality
4. Human review and validation is must
5. Create scenarios, synthetic data, and maintainable automation

Evaluating and Hardening AI-Generated Test Suites

1. AI-generated tests have strength but shallow and duplicated logic risks
2. Shallow coverage and code-tied tests don’t validate behavior
3. Human review and refactoring strengthen AI-generated suites
4. Resilient test suites focus on meaningful assertions

Preventing the AI Fix Loop

1. Fix‑prompt‑fix cycles create compounding risks in AI‑generated changes
2. Repeated AI fixes can introduce regressions and hidden breakage
3. Clear acceptance criteria and test‑first validation reduce fix‑loop churn
4. Automated quality gates help contain regression leakage
5. Confidence comes from verifying AI fixes, not blindly accepting patches

Observability-Driven Testing in AI-Enriched Systems

1. Observability provides the signals needed to validate behavior in AI‑driven systems
2. Logs, metrics, traces, and events help uncover hidden failures
3. Production telemetry enables reproducing and diagnosing AI‑generated issues
4. Observability supports root‑cause analysis and detecting behavioral regressions
5. Quality signals and SLIs guide testing of performance and non‑functional requirements

Security Validation for AI-Generated Code

1. Treat AI‑generated code as untrusted and security‑sensitive
2. AI assistants can introduce injection flaws, auth gaps, and insecure patterns
3. Common risks include insecure deserialization, secret leaks, and vulnerable dependencies
4. Threat modeling and security‑focused testing strengthen AI‑generated features

Quality Engineering Architecture for AI-Accelerated Teams

1. The test pyramid evolves to support rapid, AI‑accelerated delivery
2. Contract‑first testing and service virtualization improve stability
3. Parallel execution and ephemeral environments speed up feedback
4. Continuous testing and quality gates validate AI‑generated PRs

Maintaining Quality at Scale

1. AI‑accelerated work can create technical debt unless quality standards are defined
2. AI‑generated changes need clear review requirements and meaningful quality metrics
3. Tracking defect escapes and AI‑generated code trends reveals real quality health
4. Teams must balance productivity with quality through shared ownership
5. Sustainable governance and org‑level policies keep quality consistent at scale

Licensing, Compliance, and Organizational Policy

1. AI‑generated artifacts raise IP, licensing, and compliance risks
2. Open‑source usage requires careful license review and traceability
3. Organizations must document, audit, and govern AI‑generated changes
4. Model usage policies and regulatory guidelines shape responsible AI practice
5. Strong compliance processes prepare teams for audits and reviews

Multiple Hands-on for the topics covered

Capstone Project & Use Case

1. Project Overview
2. Complete projects to get experience and practice
3. Presentation of project findings and Insights
4. Industry Use Case Studies

Certification (Optional)

1. Certification Overview
2. Identify the right certification for you
3. Tips to prepare for certification