1 of 56

Don’ts:

No Simultaneous Hands-On: Don’t do hands-on labs during the session. Instead, follow the step-by-step guides provided.
No Interruptions: Avoid interrupting the session with unrelated queries. Off-topic questions can be asked at the end or shared via WhatsApp or sent to support@k21academy.com.

Do’s:

Ask Relevant Questions: Keep questions related to the after the session. Off-topic questions can be asked at the end or shared via WhatsApp or support@k21academy.com.
Watch Videos First: Watch the instructional videos to understand the core concepts before attempting hands-on work. Finish practice exercises in one go, and clean up immediately after practice to avoid unnecessary charges.
Create and Update Your Progress Diary: Keeping a progress diary will help you stay on track while ensuring you get the most out of the program. It’s a valuable tool for tracking your learning journey.

If you're unsure about how to set it up, please let me know.
You can find the instructions here. I’m happy to guide you through it if needed.

For Any Queries, Raise a Ticket: If you have any queries, please raise a ticket at support@k21academy.com.

Do’s & Don’t for Live Session

2 of 56

Introduction to:� AI / ML/ DL Gen AI & Agentic AI

3 of 56

Before Deployment — How AI is Built

Introduction to AI / ML / DL
Generative AI Overview
ML Pipeline Overview
Introduction to LLMs
Foundation Models (FMs) Overview

After Deployment — How AI is Used

Prompt Engineering
RAG, Vectors/Database
Chunking, Temperature, Top-p, Top-k, Hallucination
AI Agents & Agentic AI

Agenda

4 of 56

Introduction to

AI / ML / DL

5 of 56

6 of 56

Artificial Intelligence

7 of 56

Types of AI

8 of 56

Machine Learning

9 of 56

Machine Learning- 3 Types

10 of 56

Deep Learning

11 of 56

Types of Neural Networks

12 of 56

AI vs ML vs Deep Learning Comparison

13 of 56

Use Cases : AI vs ML vs Deep Learning

14 of 56

Quiz

Q1 Which of the following is the BROADEST category?

A) Machine Learning

B) Deep Learning

C) Artificial Intelligence

D) Generative AI

15 of 56

Quiz

Q2 Machine Learning is best described as:

A) Writing explicit rules for a computer to follow

B) A system that learns patterns from data without being explicitly programmed

C) A system that only works with images

D) A type of database

16 of 56

Quiz

Q3 Which type of neural network is best suited for image recognition?

A) RNN

B) LSTM

C) CNN (Convolutional Neural Network)

D) Transformer

17 of 56

Quiz Answers

1- C) Artificial Intelligence

2- B) A system that learns patterns from data without being explicitly programmed

3- C) CNN (Convolutional Neural Network)

18 of 56

Generative AI Overview

19 of 56

Generative AI

Gen AI

Model

Trained on

billions of examples

20 of 56

How Generative AI Works?

21 of 56

Generative AI — Real-World Examples

22 of 56

Generative AI vs Traditional AI

23 of 56

ML Pipeline Overview

24 of 56

ML Pipeline

25 of 56

ML Model Lifecycle

26 of 56

ML - Use Cases

27 of 56

Introduction to LLMs

28 of 56

LLM

29 of 56

LLM: key Concepts

30 of 56

How an LLM Processes Text

31 of 56

Tokenization

32 of 56

Tokenization

33 of 56

Foundation Models (FMs) Overview

34 of 56

What are Foundation Models?

35 of 56

Types of Foundation Models?

36 of 56

Prompt Engineering

37 of 56

What is Prompt Engineering?

38 of 56

Type of Prompt Engineering

39 of 56

RAG, Vectors/Database

40 of 56

RAG: Retrieval-Augmented Generation

41 of 56

RAG vs Fine-Tuning

42 of 56

RAG: Common Use Cases

43 of 56

Embeddings? - Text to Vectors

44 of 56

Vectors Databases

45 of 56

Chunking, Temperature, Top-p, Top-k, Hallucination

46 of 56

Chunking

47 of 56

Temperature: Top-p, Top-k (Controlling Output)

48 of 56

Hallucination: When AI Gets Confidently Wrong

49 of 56

AI Agents & Agentic AI

50 of 56

Agentic AI

51 of 56

Agentic AI - Use Case

52 of 56

AI Agent?

53 of 56

Agentic AI vs AI Agents

54 of 56

Complete Comparison — All 5 Technologies

55 of 56

Ask Questions

56 of 56

Thank you