1
00:00:06,150 --> 00:00:17,970
In this lesson, we are going to, uh, dive deeper into the, uh, most advanced level three LM application

2
00:00:17,970 --> 00:00:25,680
we present in this program, which is the seq insights application from Llama index.

3
00:00:27,190 --> 00:00:33,700
As always, you are going to find detailed information in the attached materials, especially in the

4
00:00:33,700 --> 00:00:34,180
notebook.

5
00:00:34,180 --> 00:00:41,950
The notebook has a lot of information about every analysis and operations we we perform in this in this

6
00:00:41,950 --> 00:00:42,580
lesson.

7
00:00:43,990 --> 00:00:48,910
So one important a note a.

8
00:00:50,420 --> 00:00:59,450
The llama index team who, as you know are is responsible from the for the development.

9
00:01:01,170 --> 00:01:03,420
Of the SEQ Insights Project.

10
00:01:04,430 --> 00:01:10,970
They have not provided a step by step guide to implement this project.

11
00:01:11,000 --> 00:01:18,920
They have provided some, uh, high level uh notes on high level explanations about, you know, the

12
00:01:18,920 --> 00:01:21,350
components and the goals of the application.

13
00:01:21,350 --> 00:01:29,660
But they have not provided a step by step a guide to reproduce the functionality of this application.

14
00:01:29,840 --> 00:01:40,880
So in order to understand the steps that they have taken, we need to use what is called reverse engineering.

15
00:01:41,000 --> 00:01:46,370
Uh, that, as you know, is not 100% precise here.

16
00:01:46,370 --> 00:01:51,800
So if you if you apply reverse engineering in the industry, for example, and I don't know, you,

17
00:01:51,800 --> 00:01:58,250
you, you, you take a piece of, I don't know, engine, for example, and you, you just, uh, break

18
00:01:58,250 --> 00:02:01,130
the, the, the engine and see the different pieces.

19
00:02:01,130 --> 00:02:09,949
You may be able to reproduce it, but in the case of a software when you have the final product.

20
00:02:09,949 --> 00:02:19,700
So the final code A, it is not easy to uh, to, to guess, uh, what has been the process to reach

21
00:02:19,700 --> 00:02:20,660
this final code.

22
00:02:20,660 --> 00:02:22,250
So you need to guess it.

23
00:02:22,250 --> 00:02:26,930
And in some cases you are going to find a problems in the way.

24
00:02:26,930 --> 00:02:35,300
So so you are going to find moments where you don't know, uh, the, the, the path to, to, to to

25
00:02:35,300 --> 00:02:35,690
follow.

26
00:02:35,690 --> 00:02:36,200
Right.

27
00:02:36,620 --> 00:02:44,120
So what we will tell you is that it is risky to try to reproduce the development process of an application

28
00:02:44,120 --> 00:02:47,390
without an step by step step guide.

29
00:02:47,630 --> 00:02:58,130
And then, in our opinion, the most, uh, useful and valuable, uh, thing you can get from this lesson

30
00:02:58,130 --> 00:03:00,230
is to get the big picture.

31
00:03:01,090 --> 00:03:08,890
So to understand some details about, you know, the process and about the the components and about

32
00:03:09,220 --> 00:03:15,130
the way that this application seems to be built, but do not.

33
00:03:15,130 --> 00:03:22,990
Well, I will I will recommend you to be careful when trying to reproduce the development process of

34
00:03:22,990 --> 00:03:29,740
this project, because we don't have enough information, but we have a lot of, uh, data.

35
00:03:29,740 --> 00:03:33,460
So we are going to do our best with the data we have.

36
00:03:34,180 --> 00:03:39,970
You will see that a in the a practical video.

37
00:03:40,000 --> 00:03:40,780
We will see.

38
00:03:40,780 --> 00:03:46,090
Next we are going to highlight some details about this application.

39
00:03:46,090 --> 00:03:51,070
For example, it is very important for you to use GitHub code space.

40
00:03:51,220 --> 00:03:54,760
Uh in order to work with this project.

41
00:03:54,760 --> 00:03:57,580
They provide two different options.

42
00:03:57,580 --> 00:04:01,810
And in my opinion, the other option is not a good idea.

43
00:04:02,200 --> 00:04:05,920
And then we are going to talk about the tech stack in detail.

44
00:04:05,920 --> 00:04:14,380
The LM logic that the the application is using, the external APIs and the complexity level they are

45
00:04:14,380 --> 00:04:14,950
adding.

46
00:04:15,580 --> 00:04:24,340
So from one side, this is interesting to have a, uh, you know, external APIs working with your application.

47
00:04:24,340 --> 00:04:28,300
But in this case they are complex, uh, APIs.

48
00:04:28,300 --> 00:04:34,480
And if you want to reproduce the functionality and use this external APIs is going to be a little bit

49
00:04:34,480 --> 00:04:35,410
cumbersome.

50
00:04:35,410 --> 00:04:41,080
So we will see this and other details that I think is are going to be very interesting for you at this

51
00:04:41,080 --> 00:04:42,370
moment of the program.

52
00:04:42,460 --> 00:04:51,700
And when we finish with this, we are going to review, uh, two other interesting level three applications

53
00:04:51,700 --> 00:04:58,090
that we think are very, very good for you at this moment in the program.

54
00:04:58,090 --> 00:04:58,450
Okay.

55
00:04:58,450 --> 00:05:07,780
So first let's go into the deep dive of the SEQ Insights Project in the next practical video.