Full Transcript

·YouTLDR

You need to Learn Docker NOW! - It isn't Just for Cloud or DevOps Anymore

12:24EnglishTranscribed Jul 23, 2026
0:00

There is a huge misconception that if

0:03

you want to specialize in AI, machine

0:06

learning, data science or become a data

0:08

analyst, data scientist, cyber security

0:11

professional and all that then you can

0:13

forget about the infrastructure and the

0:15

infrastructure skills. So basically

0:17

forget about the cloud, forget about

0:19

automation tools and all that which is

0:20

wrong. In this video I will show you why

0:24

you need to learn Docker and why you

0:26

have to learn it. Now, if you are in IT

0:29

or you are considering to be in IT in

0:32

the near future or you would like to

0:33

have a lot of tools in your skill set

0:37

box that you can use and leverage. So,

0:40

let's go ahead and ask Chad

0:44

GPT. So, let's go ahead and do a simple

0:46

search on one of the AI tools available.

0:50

Why is Docker valuable and beneficial

0:55

to a lot of professions that we are

0:58

going to

1:00

discuss. But before that, please

1:02

subscribe to the channel. There will be

1:04

a lot of videos coming up in English

1:06

that are going to be beneficial to

1:08

anyone who wants to be in it in the

1:11

foreseeable future. So, you don't want

1:12

to miss anything. activate the

1:14

notification, give us a like and share,

1:16

and let the video reach as much people

1:19

as possible to benefit everyone. So,

1:21

let's go ahead and start with what is

1:23

Docker? Why is Docker so appealing and

1:26

so attractive right now to simplify the

1:29

matter, we have software and the

1:32

software needs computations or computing

1:34

devices, computing power to run, right?

1:37

So that could be on a server on a

1:39

physical server. It could be on a

1:40

virtual machine VMware, HyperV and so on

1:43

or it could be on containers or it could

1:47

be containerized. It could be run in

1:49

software. Basically physical servers and

1:52

virtual machines they have issues.

1:54

issues like what lack of optimized usage

1:57

of resources, slow or long time to

2:01

start, slow start um the the fact that

2:04

when it is ported to other environments,

2:07

it's not guaranteed it will work without

2:09

problems. Physical servers, you have to

2:11

shift and you have to lift and shift the

2:13

server in order to get to another place.

2:15

The beauty of containers is the fact

2:18

that it's portable, it's highly

2:20

scalable, it's very fast to boot and all

2:22

that. So let's find out in layman terms

2:24

what is a container. So as you can see

2:28

on the screen a container we are going

2:30

to visualize that as a box. The box

2:33

could be as you can see cardboard or it

2:36

could be the container the shipping

2:38

container on the

2:40

ships. So what is the idea of

2:42

containers? You need to put everything

2:44

that your application or your code needs

2:47

to run in one place and lock it.

2:51

So you put your code be it Python, be it

2:53

Java, R whatever it is and whether you

2:56

are working on machine learning on cyber

2:58

security or you are a developer, you are

3:00

doing automation, you are a devops

3:02

engineer, a data scientist, MLOps

3:04

engineer, all of that you will need to

3:06

write code to do what whatever is

3:09

supposed to

3:10

happen and the code needs an interpreter

3:13

so the computing device or the computing

3:15

machine can understand it. That's why we

3:17

need the runtime. For example, if this

3:19

is Python, we need let's say Python 3

3:22

for example to be

3:24

installed. Then there would be

3:27

configuration parameters like which

3:29

version, what is the location of this,

3:31

what is the URL of the database you

3:33

would like to connect to. All of that is

3:35

going to be included as well in the box

3:38

in the container. And then you add any

3:41

dependencies. You need this driver. You

3:43

need that version of this, that version

3:45

of that. All of these the passwords

3:48

environment variables that you need all

3:50

of that is going to be contained and

3:52

then we seal the box we seal the

3:55

container. So once we do this then all

3:58

what we need is to transform that into

4:01

an image a template from which we can

4:03

run the actual containers. So what is a

4:06

container then? The container is nothing

4:08

more than a program, a software program

4:11

that runs as an isolated process in your

4:15

environment. So what's my environment?

4:18

That could be a physical server, it

4:19

could be a virtual machine. So a

4:20

software component that runs on top of

4:24

your computing device. If you noticed

4:28

here, we haven't mentioned anything

4:29

about operating systems. We didn't say

4:31

that we're going to add Windows or Linux

4:33

or Mac OS or Ubuntu or Real or whatever

4:36

the operating system required to run

4:38

this. We don't you don't include that in

4:41

the container and therefore the features

4:44

of the containers that make it very

4:47

appealing is it doesn't require a

4:49

separate OS. So the size is very small.

4:51

You don't need that 40 or 70 gigs for

4:54

the operating system to execute and run.

4:58

It's isolated. Once we seal it, it

5:00

becomes like the shipping shipping

5:02

container on the back of the boat or on

5:05

the docks of the boats. It doesn't know

5:07

anything about the containers around.

5:09

The containers around don't know

5:10

anything about the content of these

5:12

container. So, it's all self-contained

5:14

environments and self-sufficient because

5:16

they have everything they need to run

5:19

because it's small. It's very fast to

5:21

run. It's very fast to tear down. We're

5:22

talking about milliseconds in this

5:25

case. It is repeatable. use the same

5:28

template to run container anywhere on

5:30

AWS, on Google, on Azure, on premises,

5:32

it's going to be the same result and

5:35

definitely that also makes it portable.

5:37

So today I'm with AWS. I'm not very

5:40

happy with AWS. I'm going to ship my

5:41

templates elsewhere and I'm going to

5:43

move my production workloads elsewhere

5:45

as well. And it's easily scalable

5:48

software component 60 megs, 60

5:51

megabytes, 100 megabytes, 200 megabytes,

5:53

half a gig. very easy to scale because

5:56

it spins in milliseconds. So this is

5:59

very important. But how does this magic

6:01

happen and how can I run the code on a

6:03

computing device without an operating

6:04

system? Doesn't make sense, right? So

6:06

let's find

6:08

out. So to run the containers, you need

6:11

a host and the host can be a virtual

6:14

machine or it can be a physical server.

6:19

On the host there will be the hardware

6:21

required the CPUs the RAM the disks and

6:24

the G GCPUs and all that. There is an

6:27

operating system that is going to be

6:28

shared among the different containers

6:31

and the mechanism to share that is

6:33

through the engine. This is the

6:36

mastermind. This is the brain of the

6:38

dockerized or the

6:40

containerization. And this is how the

6:43

different containers they access the

6:45

kernel of your operating system on the

6:47

host as if they had their own operating

6:49

system. So that's how we succeed in

6:52

making

6:53

this work this way. All right. So we

6:56

know this. Now let's go back and use AI

6:59

to tell us how is this beneficial

7:03

containerization or docker and learning

7:05

docker is beneficial in the different

7:07

professions that you could be aspiring

7:09

or maybe are working on right now. So

7:12

here we go. I asked Chad GPT, would

7:15

Docker be beneficial to data analysts,

7:18

machine learning engineers, AI

7:19

engineers, data scientists, and MLOps

7:22

engineers? And the answer was with

7:26

examples, Docker is about productivity

7:30

boosting, major productivity boost. And

7:33

that is why no matter what your

7:35

application is, no matter what your use

7:37

is, you will need to use that tool to

7:41

boost your productivity. So let's take

7:43

it with data analysts. The benefit is it

7:46

will ensure consistent environments.

7:49

Whether you're using Jupyter Notebooks,

7:51

Python, R or any database clients, it's

7:54

going to provide you consistency

7:56

wherever you work and whenever you work.

7:59

And here's a use case. running a local

8:02

container with a Jupyter notebook that

8:04

already has pandas, numpy and other

8:08

libraries in Python and a connection to

8:10

Postgress or Snowflake. So here's an

8:13

example that where the data analyst if

8:16

he or she they know how to use Docker

8:19

then they can be very productive and

8:21

they can be very fast to deploy whatever

8:23

they are working on as projects or to

8:25

test it. Machine learning engineers

8:27

benefit easily isolate model training

8:30

environments. You need isolation. So in

8:32

that case whether you're using

8:34

TensorFlow,

8:35

PyTorch, you need to do that isolation

8:38

and you need to scale without breaking

8:40

the

8:42

actual runtime that you are working on.

8:45

Use case run GCPU or GPU enabled

8:48

training in Docker container using

8:50

Nvidia Docker. AI engineers. It will

8:54

help them build and deploy deep learning

8:56

models with specific dependencies like

8:58

the O

8:59

andNX or hugging face

9:03

and packaging those inside containers.

9:07

And here is a use case as well for that

9:09

for inference services with fast API or

9:12

Flask

9:16

plus inside containers and ship it to

9:19

staging or to production environments.

9:20

Then you have data scientists share

9:23

research easily whether you are using uh

9:26

full experiments with notebooks and so

9:28

on and dependencies they can be

9:29

contained and then you can build the

9:31

image template and then send it and

9:32

share it with others. So this is

9:34

reproducible experimentation MLOps

9:37

engineers machine learning

9:39

ops repeatable automated and scalable

9:43

deployments is the core of MLOps like in

9:46

DevOps. So Docker is the core of that.

9:49

And here's an example. Model service uh

9:52

pipelines, CI/CD workflows and mon

9:55

monitoring tools all

9:57

containerized. So the common thing

10:00

across all IT professions that Docker is

10:02

going to provide you is environmental

10:05

reproducibility. You can reproduce the

10:07

environment in development, in testing,

10:09

in staging on AWS or elsewhere.

10:11

Dependency management. So now it's all

10:14

self-contained within that template or

10:16

image. So you don't need to worry about

10:18

conflicts between different dependencies

10:20

for different applications you're

10:23

using. Portability, the fact that you

10:26

can run it anywhere you would like on

10:28

your development environment, on your

10:29

laptop, in the cloud, on premises, and

10:32

version control because it's all about

10:35

scripts and the scripts can be version

10:38

controlled. So this tells you why no

10:40

matter what the profession that you are

10:43

aspiring from the hot skills in IT, you

10:46

have to learn Docker. Of course, I did

10:47

not include DevOps. I did not include

10:50

development. I did not include cyber

10:52

security. I did not include cloud

10:54

because it's obvious. It's obvious that

10:56

they are needed today. This is known.

10:58

What is not known is the assumption or

11:00

the the misconception that if I'm going

11:03

to be in AI or machine learning or data

11:05

science then I don't need really really

11:07

need to worry about cloud or uh docker

11:10

specifically for this one. Of course,

11:13

when you scale the containerized usage

11:16

or the containerized application

11:18

efforts, then going into Kubernetes is

11:21

going to be the right step to do. And

11:23

probably I'll make another video also

11:25

the same style to explain to you why do

11:27

you need to do that. So learn Docker and

11:31

learn it right now because it's going to

11:34

help you master your job no matter what

11:36

that is and no matter what your career

11:39

path is within it. If you are a

11:41

newcomer, if you are transforming into

11:43

IT or if you are upskilling into IT,

11:46

learn Docker if you haven't done so and

11:48

learn it right now. And it's very very

11:51

easy to learn. I'm working on a course

11:54

that will be published in the next few

11:55

days in English and that will take you

11:57

from zero to pro in Docker very fast. So

12:02

look in the description box for the URL

12:04

and I will see you in the next video.

12:06

But please consider subscribing to the

12:09

channel. activating the notifications

12:12

and give us a like and share. Let the

12:14

video reach as many people as possible

12:17

to benefit immensely. Thank you so much

12:20

and I will see you in the next video.

More transcripts

Explore other videos transcribed with YouTLDR.

Get the TLDR of any YouTube video

Transcribe, summarize, and repurpose videos in 125+ languages — free, no signup required.

Try YouTLDR Free