1
00:00:05,110 --> 00:00:11,320
In this lesson, we will talk about the advanced architecture of an LM application.

2
00:00:17,000 --> 00:00:24,590
Remember that in the previous lesson we talked about the toy demos, what we call the basic architecture

3
00:00:24,590 --> 00:00:26,360
of Elm applications.

4
00:00:26,360 --> 00:00:34,490
And if you remember in this program, we talk about level one applications, when we also have when

5
00:00:34,490 --> 00:00:40,370
we only have three components Elm Orchestration Framework and Vector Database.

6
00:00:40,370 --> 00:00:43,790
And we talk about a level two applications.

7
00:00:43,790 --> 00:00:51,530
When a we have these three components plus a what we call a toy user interface.

8
00:00:52,310 --> 00:00:52,880
Case.

9
00:00:52,880 --> 00:00:59,180
So the level one application and the level two applications are toy demos for us.

10
00:00:59,180 --> 00:01:01,880
These are not professional applications.

11
00:01:01,880 --> 00:01:08,300
When we are talking about a professional applications and we are talking about a full stack application,

12
00:01:08,300 --> 00:01:14,120
what we call level three application or professional level application.

13
00:01:14,120 --> 00:01:17,840
In this case, we have a back end framework.

14
00:01:18,660 --> 00:01:20,700
Uh, front end framework.

15
00:01:20,850 --> 00:01:22,200
A back end server.

16
00:01:22,230 --> 00:01:23,850
A front end server.

17
00:01:23,850 --> 00:01:28,830
And then we have external APIs and cloud storage.

18
00:01:29,040 --> 00:01:32,340
Uh, plus the LLM foundational model.

19
00:01:32,340 --> 00:01:37,710
So the most typical advanced architecture we are going to find.

20
00:01:38,550 --> 00:01:42,990
Will be similar to the one you have in this slide.

21
00:01:43,020 --> 00:01:46,590
So for back end server we will use render.

22
00:01:46,590 --> 00:01:48,810
Com a.

23
00:01:48,810 --> 00:01:56,190
For back end framework we will use fast API which is a framework based on the Python language.

24
00:01:56,280 --> 00:02:03,090
We will use an orchestration framework that can be long chain or lambda index.

25
00:02:03,090 --> 00:02:12,420
And we will use one or more than one databases, usually a one vector database and one traditional database.

26
00:02:13,520 --> 00:02:26,960
In the front end a framework we are going to use Next.js, which is a type of react, a framework and

27
00:02:26,960 --> 00:02:30,560
a we will have Vercel.

28
00:02:30,560 --> 00:02:33,770
Com as a front end server.

29
00:02:34,700 --> 00:02:35,420
Next year.

30
00:02:35,420 --> 00:02:38,030
Yes, is based on JavaScript.

31
00:02:38,030 --> 00:02:42,560
So in the backend we are going to play around with Python.

32
00:02:42,560 --> 00:02:45,710
And in the front end we are going to play with JavaScript.

33
00:02:45,710 --> 00:02:51,680
Don't worry, it will be easier and more simple than what you may think.

34
00:02:51,710 --> 00:02:55,970
Then we will have a external APIs.

35
00:02:56,770 --> 00:03:01,780
Connected with the back end framework, also the LM model.

36
00:03:01,960 --> 00:03:13,900
Usually ChatGPT, and in many cases we are going to use cloud storage to store a our private files like

37
00:03:13,900 --> 00:03:15,520
PDFs, for example.

38
00:03:15,520 --> 00:03:22,150
And in many cases we are going to use S3 from Amazon Web Services.

39
00:03:27,490 --> 00:03:37,060
So as you can see, the difference between a basic architecture and an advanced architecture is important.

40
00:03:37,420 --> 00:03:41,380
A these are much more complex solutions.

41
00:03:41,380 --> 00:03:50,320
And you need to understand that besides this architecture, we will need to work with a validation frameworks,

42
00:03:50,320 --> 00:04:00,460
with LM ops, with a LM cache, etc., etc. we will examine all these elements in detail in the program.

43
00:04:03,030 --> 00:04:10,890
In the next lesson, we are going to preview a professional LM application.