WEBVTT
1
00:00:10.269 --> 00:00:13.470
Hi, I'm your hus premier'll be
sunker, and this is the data standard
2
00:00:13.470 --> 00:00:17.030
audio experience. We have a very
exciting guest today on our podcast. We
3
00:00:17.109 --> 00:00:21.859
have anthony a Kassatti, data scientists
and global search strategies. Today we're speaking
4
00:00:21.859 --> 00:00:26.780
about his journey into data science,
as well as applications within small beating manufacturing.
5
00:00:27.140 --> 00:00:29.300
Welcome to the show. Thank you. Thank you for having me.
6
00:00:29.579 --> 00:00:32.619
Yeah, of course, thanks for
coming on to the show today. Go
7
00:00:32.740 --> 00:00:35.969
ahead and introduce yourself. Sure so, like I said in, my name
8
00:00:36.009 --> 00:00:39.530
is anthony a Kastadi. I am
an independent data scientist and I work with
9
00:00:39.929 --> 00:00:45.450
a lot of marketing teams and data
science teams at various manufacturers or technology companies
10
00:00:45.490 --> 00:00:51.560
and healthcare destroy myself to create a
new methodologies and technologies to help them kind
11
00:00:51.600 --> 00:00:55.600
of like take a step further in
the end others. Yeah, definitely.
12
00:00:55.640 --> 00:00:59.719
Well, I'm really excited to have
you for this conversation today, and your
13
00:00:59.799 --> 00:01:03.439
career thus far has been incredibly impressive, with you first starting as an electrical
14
00:01:03.520 --> 00:01:07.709
engineer and then making the transition to
like an independent data science consultant, as
15
00:01:07.750 --> 00:01:10.790
you are now. How did you
end up working in data science and what
16
00:01:10.909 --> 00:01:14.189
has been the hardest part of transition. So yeah, I mean I mean
17
00:01:14.629 --> 00:01:18.269
for the past ten to eleven years
it never occurred to me that I'll be
18
00:01:18.469 --> 00:01:22.579
here like as a data scientist,
and it was not even a job back
19
00:01:22.620 --> 00:01:26.180
then, right. So when I
was finishing my engineering I had to look
20
00:01:26.180 --> 00:01:30.620
for work and at that time there
was a boom in terms of like searching
21
00:01:30.739 --> 00:01:36.569
and optimization and they were people selling
services online and kind of like. That
22
00:01:36.849 --> 00:01:42.489
increased the visibility for web developers and
software engineers to build new applications and everything.
23
00:01:42.530 --> 00:01:47.010
So that's how I went entered into
the web development field. So I
24
00:01:47.090 --> 00:01:49.120
worked at a marketing agency in Melbourne. Could the good thing is and in
25
00:01:49.239 --> 00:01:53.560
engineering you actually are you learn.
It's a multi faceted course. You you
26
00:01:53.640 --> 00:01:57.480
learn computer science, you'll learn engineering, you also learn a little bit of
27
00:01:57.599 --> 00:02:01.790
business administration and everything in there.
So the this part of me always like
28
00:02:01.989 --> 00:02:07.189
programming, mathematics and everything. So
as a developer I was able to kind
29
00:02:07.189 --> 00:02:09.550
of like get into like the data
collection process. I was able to build
30
00:02:09.550 --> 00:02:14.870
multiple software applications and everything and slowly
over the course of ten years, I
31
00:02:14.949 --> 00:02:19.219
mean you you ended up being a
data signs them. Yeah, it's a
32
00:02:19.300 --> 00:02:22.860
cool it's a cool thing to have, like your data signers, but it's
33
00:02:22.860 --> 00:02:24.580
also like a cool job as well
then. Yeah, yeah, definitely.
34
00:02:24.979 --> 00:02:30.129
And what's been like, at least
for you? The hardest part about being
35
00:02:30.169 --> 00:02:35.370
a data scientists? I mean the
hardest part is it is so to keep
36
00:02:35.370 --> 00:02:39.770
it to keep it simple. You
there are multiple different segments within the data
37
00:02:39.810 --> 00:02:44.090
signs, right. So when you
branch out a lip, step back a
38
00:02:44.090 --> 00:02:46.159
little bit, you look at Ai
am all in deep learning, and you
39
00:02:46.680 --> 00:02:53.719
the software engineers, database administrators or
people who are using people who are kind
40
00:02:53.719 --> 00:02:57.000
of like doing the data mine stuff. Now it broken down into like architects,
41
00:02:57.080 --> 00:03:00.310
engineers and data signed this kind of
thing. So the hardest part is
42
00:03:00.349 --> 00:03:02.229
to kind of like you find your
fit, right, to find your fit
43
00:03:02.310 --> 00:03:07.509
in an organization. So, if
you're trying to transition, do you want
44
00:03:07.509 --> 00:03:10.030
to be data scientists? Do you
want it to I mean say if you
45
00:03:10.110 --> 00:03:14.379
want to, if you're an engineer
or an analyst, that there are certain
46
00:03:14.460 --> 00:03:17.219
things that you always do want a
day to be basis right and you don't
47
00:03:17.259 --> 00:03:21.780
want to kind of like be there
for a longer time because they are to
48
00:03:21.860 --> 00:03:24.219
science. Is We're growing. I
mean in the last five years itself,
49
00:03:24.860 --> 00:03:30.250
we've outgrown ourselves in terms of writing
algorithms and everything, and basically we are
50
00:03:30.409 --> 00:03:35.569
making ourselves out of business right,
because that you now have. You not
51
00:03:35.729 --> 00:03:40.719
have the tools and libraries where you
can run thousands of algorithms and ten minutes.
52
00:03:42.120 --> 00:03:45.039
Back in the day, I mean
like Justin like last three years,
53
00:03:45.159 --> 00:03:47.360
I actually took probably a day or
two to kind of like write an algorithm.
54
00:03:47.479 --> 00:03:52.319
Right. Normally arena regulation not came. Came means clustering or something like
55
00:03:52.360 --> 00:03:53.960
that. Now it's like you have
ten different things that you can do at
56
00:03:53.960 --> 00:03:57.509
the same time. So I think
the hardest part is to wear is to
57
00:03:57.590 --> 00:03:59.669
kind of like you find your fit. What do you want to kind of
58
00:03:59.710 --> 00:04:01.389
like move forward? And where are
you doing that? I mean you are
59
00:04:01.389 --> 00:04:04.069
you doing at a a small scale, or you're doing at a bigger scale?
60
00:04:04.069 --> 00:04:10.379
Are you working at the Amazons,
the Google kind of like space,
61
00:04:10.460 --> 00:04:14.979
or are you actually trying to kind
of like improve someone's business operations, improve
62
00:04:15.060 --> 00:04:17.579
customer experiences or something like that?
So that's kind of like the way thought
63
00:04:17.620 --> 00:04:20.379
process that you need to have.
Huh, that's interesting. So like finding
64
00:04:20.420 --> 00:04:25.490
your niche is really important within data
science. Yeah, every day that,
65
00:04:25.529 --> 00:04:29.490
I mean not all day is very
common, and even even the volume and
66
00:04:29.529 --> 00:04:32.209
the wallacity at which the data comes
in is also not common. So you
67
00:04:32.610 --> 00:04:36.970
never can actually work at a small
scale company, go directly to like an
68
00:04:36.970 --> 00:04:43.079
Amazon kind of service and figure out
that I can. So you're always bombarded
69
00:04:43.160 --> 00:04:46.120
with something very new and something really
exciting. Always yeah, well, that's
70
00:04:46.160 --> 00:04:51.040
really cool. And do you think
coming from that background helped you anyway,
71
00:04:51.319 --> 00:04:57.509
or it was it just like what
you learned within the within your way as
72
00:04:57.550 --> 00:05:00.829
a becoming a data scientist? I
me yes, it absolutely does. I
73
00:05:00.949 --> 00:05:06.699
mean having an engineering background absolutely helps
in transitioning to the data signs we are
74
00:05:08.379 --> 00:05:13.019
but I would not say I would
probably recommend it, because I have friends
75
00:05:13.060 --> 00:05:16.300
and family and also colleagues who come
from different verticals, who come from accounting.
76
00:05:16.699 --> 00:05:20.569
Our data analysts, right, who
are who done the hut on psychology,
77
00:05:20.649 --> 00:05:24.810
and they are also kind of like
there aren't analyzing customer behavior, and
78
00:05:24.850 --> 00:05:28.410
also they are, there are a
type of data assigned to themselves. I
79
00:05:28.529 --> 00:05:30.889
mean, if you're understanding the code
of data science, you need to always
80
00:05:30.970 --> 00:05:36.319
have an interest in math rate because
you're more of a statistician than software engineer.
81
00:05:36.399 --> 00:05:41.199
You're better right, if you're better
like that. So you think of
82
00:05:41.279 --> 00:05:44.000
it in that way and depending on
the job as well, you if you
83
00:05:44.240 --> 00:05:46.079
look at the job descriptions or the
companies to kind of like hire you.
84
00:05:46.279 --> 00:05:50.269
Sometimes they ask you for like an
engineering but some who you ask you for
85
00:05:50.509 --> 00:05:54.829
like a PhD in statistics of instead
physics or something like that. So it's
86
00:05:54.829 --> 00:05:58.509
always about the job. where,
where are you going in and where are
87
00:05:58.550 --> 00:06:01.269
you kind of like placing our so
and engineering background always helps. I mean
88
00:06:01.310 --> 00:06:06.980
that's how it helped me transition from
a completely filled job like an engineering like
89
00:06:08.100 --> 00:06:10.699
you go out on the field,
get the job done and everything. But
90
00:06:10.939 --> 00:06:15.060
for me it eventually turned out that
I like analytics, I like rip,
91
00:06:15.459 --> 00:06:17.850
I like how numbers right. I
have to make narratives out of it,
92
00:06:18.009 --> 00:06:20.569
meaning out of it, get tell
story out of it. So started to
93
00:06:20.610 --> 00:06:25.170
help businesses, started to see the
change. Obviously, as an engineer,
94
00:06:25.250 --> 00:06:28.250
I wanted to do something and I
wanted to see the effects of it so
95
00:06:28.449 --> 00:06:31.290
I can have feel pleasant right,
I mean there's something in the world that
96
00:06:31.329 --> 00:06:35.160
I'm actually doing, that it's actually
working. So it happens in data signs,
97
00:06:35.160 --> 00:06:40.279
I mean you can now that everything
is instant in real time, with
98
00:06:40.600 --> 00:06:44.079
a lot of data that's coming in. I mean you don't necessarily have any
99
00:06:44.199 --> 00:06:47.269
filters everywhere. So it is a
truly good experience that you can have with
100
00:06:47.430 --> 00:06:51.470
this one. I mean, even
though you compaplate. I mean your story
101
00:06:51.550 --> 00:06:55.709
is really cool because a lot of
our listeners are students that are trying to
102
00:06:55.750 --> 00:06:59.470
break into data scientists, or at
least people that are within different industries that
103
00:06:59.509 --> 00:07:01.500
are trying to, you know,
get into data science because they see that
104
00:07:01.579 --> 00:07:05.660
it's a like a really cool industry. So for those people, do you
105
00:07:05.819 --> 00:07:11.139
have any advice or anything that,
like when you first came to the industry,
106
00:07:11.139 --> 00:07:14.500
where there any misconceptions that she had
to like get over that yourself,
107
00:07:14.740 --> 00:07:16.930
or any advice that you'd give to
people such as that trying to break into
108
00:07:16.930 --> 00:07:20.370
the industry? Yeah, I think
that in my early days, we're a
109
00:07:21.009 --> 00:07:25.370
C shop. Java had do big
data. This was kind of like a
110
00:07:25.930 --> 00:07:30.959
big, big thing, and so
you start to see companies, recruiters and
111
00:07:30.600 --> 00:07:33.480
people who hire you kind of like
ask you for everything, right, I
112
00:07:33.879 --> 00:07:36.079
you could a job, or you
could at C plus ps, I could
113
00:07:36.120 --> 00:07:40.480
at Python, I mean anything that
comes in right. So the misconception is
114
00:07:40.800 --> 00:07:44.279
be good at something, don't just
for the sake of it, if there's
115
00:07:44.279 --> 00:07:46.069
something coming in the industry or if
they if there's a hype about it,
116
00:07:46.230 --> 00:07:49.389
if there's a new programming language that
you have to learn for it, and
117
00:07:49.589 --> 00:07:53.949
probably Yah who's written something, or
a Google as as brought in some code.
118
00:07:54.350 --> 00:07:56.910
You don't necessarily have to kind of
like jump in right away about it,
119
00:07:56.990 --> 00:08:00.259
but just do your research on it, look at how that affects your
120
00:08:00.459 --> 00:08:03.420
current job and look at your career
progression as well, kind of like go
121
00:08:03.660 --> 00:08:07.980
through the forums and everything. But
the way that I was trying, I
122
00:08:07.060 --> 00:08:11.180
mean I was miss come misconceptualized,
is I have to learn everything after be
123
00:08:11.259 --> 00:08:13.649
good at good, be at good, be good at everything so I can
124
00:08:13.970 --> 00:08:16.769
take the interviews properly, get to
get the job done and everything like that.
125
00:08:16.970 --> 00:08:20.649
But it's always it's the company's choice
to kind of like hire you or
126
00:08:20.689 --> 00:08:24.089
not. Right. You may be
good at our or you may be good
127
00:08:24.129 --> 00:08:26.970
at the Python, but maybe the
company can come and ask I only need
128
00:08:26.089 --> 00:08:28.800
people for Jova. That doesn't make
Oud. You actually have to be like
129
00:08:30.000 --> 00:08:33.879
hundred percent expert in Java like that. So think about it, whether it's
130
00:08:33.879 --> 00:08:37.279
a right choice or not, and
they don't owe stress yourself and data sense.
131
00:08:37.360 --> 00:08:39.799
Is Not a job to get stressed. It's a cool job. You
132
00:08:39.919 --> 00:08:43.669
have to keep your brain pleasant because
you'll have to keep thinking always, because
133
00:08:43.990 --> 00:08:48.549
you have people come to you with
their requirements, with their problems. You
134
00:08:48.789 --> 00:08:50.750
probably need to find a solution with
those numbers. Rate, you are the
135
00:08:50.830 --> 00:08:56.220
only person that they rely on to
get the expert feedback one right. So
136
00:08:56.500 --> 00:08:58.820
it's not always that you have a
magic had and you take a rye bit
137
00:08:58.899 --> 00:09:01.700
out of it, but it's always
kind of like you have to kind of
138
00:09:01.740 --> 00:09:05.460
like make sense out of these numbers, and that's why they rely on you.
139
00:09:05.700 --> 00:09:09.259
Right. So I would say that's
kind of like the first advice I
140
00:09:09.299 --> 00:09:11.090
would like, the first in the
last eddy I would give you that will
141
00:09:11.169 --> 00:09:16.730
actually shape you to being a better
person at me taking making decisions and taking
142
00:09:16.809 --> 00:09:20.090
decisions. That will also give you
a effect in terms of how you work
143
00:09:20.210 --> 00:09:24.210
in your workplace as well. Hey, yeah, thank you. That makes
144
00:09:24.360 --> 00:09:28.080
a lot of sense. I actually
wanted to switch over from like your pastory
145
00:09:28.240 --> 00:09:31.799
to like the work that you're currently
doing. I understand that you mostly deal
146
00:09:31.879 --> 00:09:35.440
with small menium manufactures and, as
someone who doesn't really know too much about
147
00:09:35.440 --> 00:09:39.629
the industry, would you mind explaining
the biggest challenges that you face and also,
148
00:09:39.909 --> 00:09:43.549
like the word that you do on
the clients that you work with.
149
00:09:43.950 --> 00:09:46.549
Yeah, sure, sure, yes, small menuon manufacturers is one of the
150
00:09:46.590 --> 00:09:50.590
niche that I ply worked in the
experience that I had in the US,
151
00:09:50.789 --> 00:09:54.379
and I also worked with less than
fifty million dollar raven new companies in here
152
00:09:54.539 --> 00:10:00.059
are and I also work with smaller
companies as well, right, I mean
153
00:10:00.100 --> 00:10:03.419
these are companies that are attacked to
consumer. You to see kind of department
154
00:10:05.179 --> 00:10:07.289
like a background. Out of all
of this, I figured one thing that
155
00:10:07.450 --> 00:10:11.090
is data everywhere that the company necessarily
kind of like consumes, right, so
156
00:10:11.610 --> 00:10:15.450
the type of their time and how
they consume is not necessarily thought of.
157
00:10:16.169 --> 00:10:20.049
So you just have lots and lots
of volume coming in, but you never
158
00:10:20.169 --> 00:10:26.519
necessarily understood how to kind of like
strategize structure it for your consumption and how
159
00:10:26.559 --> 00:10:28.679
do you want it to take up
learn, like the build, test and
160
00:10:28.799 --> 00:10:33.480
learning model is prop is not properly
thought of. So what I've done basically
161
00:10:33.639 --> 00:10:37.350
was to kind of like come in
and you have a lot. You have
162
00:10:37.389 --> 00:10:39.429
a lot of content out there for
data science. You have a lot of
163
00:10:39.470 --> 00:10:41.830
youtube videos, you a lot of
people talking about it. You have a
164
00:10:41.870 --> 00:10:46.590
lot of food cams and exercises that
people can take to learn theory tically what
165
00:10:46.789 --> 00:10:48.909
their assience is. And then you, if you wanted to kind of implement
166
00:10:48.950 --> 00:10:54.860
that in your workplace. You probably
may not necessarily take a cost effective cost
167
00:10:54.940 --> 00:11:00.019
division to decision. Right, you
probably are choosing for tools that are beyond
168
00:11:00.139 --> 00:11:05.250
what you need. Right. Yeah, it's just like you using Microsoft word,
169
00:11:05.649 --> 00:11:07.970
but you only use like a few
tools, and microsoftware, you don't
170
00:11:07.970 --> 00:11:11.649
use its whole spectrum of things.
They're right. So, so what I've
171
00:11:11.690 --> 00:11:16.129
done basically is to kind of like
it's if you go in, you understand
172
00:11:16.129 --> 00:11:20.639
the business needs, right. So
most of these companies they have transactional data,
173
00:11:20.720 --> 00:11:24.440
they have customer behavior data, customer
care data that coming from their phone
174
00:11:24.480 --> 00:11:28.600
calls or the contact form their inquiries. Some has warranty issues, some may
175
00:11:28.679 --> 00:11:33.350
have product problems, right, so
all of this has to be mind in
176
00:11:33.389 --> 00:11:37.070
terms of understanding to innovate your product. Basically, what would be my next
177
00:11:37.149 --> 00:11:43.389
iteration rate? And that's where I
come in when I set up master data
178
00:11:43.429 --> 00:11:50.100
management pipelines where we create ways in
which we channelize these fountains of information into
179
00:11:50.259 --> 00:11:54.899
silos, right, and depending on
different teams you may have anop you and
180
00:11:54.100 --> 00:11:56.820
operations team, you have Ay,
you may have an e commerce team,
181
00:11:56.820 --> 00:12:01.889
you have have pricing teams, you
have engineering teams that you always have to
182
00:12:01.970 --> 00:12:07.210
feed information, right. That's how
they take informed decisions and everyone is a
183
00:12:07.330 --> 00:12:09.409
part of the state ASIGENCE team.
It's not like dat designers is only kind
184
00:12:09.409 --> 00:12:13.850
of liquing everything, right. So
I kind of like bring that culture in
185
00:12:15.330 --> 00:12:20.879
as well and for company he's who
never had these experiences. Obviously a shift
186
00:12:20.960 --> 00:12:26.879
right in an agile way of doing
things. So it's obviously a practice that
187
00:12:26.919 --> 00:12:30.710
you'll have to keep doing it and
show the outcome of it. And and
188
00:12:31.549 --> 00:12:35.230
companies are actually going trying to take
that transition as well. So you don't
189
00:12:35.230 --> 00:12:39.070
need to invest heavily on data signers
event there's a misconstruption, a misconception within
190
00:12:39.149 --> 00:12:43.190
the organizations where you need get aware
of houses. You need etl tools,
191
00:12:43.269 --> 00:12:48.019
you probably need tools and tool sets
and services. You any licenses and and
192
00:12:48.500 --> 00:12:52.460
it's an anyways. It's an up
upwards to a million to like couple of
193
00:12:52.539 --> 00:12:56.539
million dollars a year in terms of
like the budget that the company makes.
194
00:12:56.580 --> 00:12:58.610
But I don't necessarily feel that is
important. I mean, at the moment,
195
00:12:58.730 --> 00:13:01.529
you can actually do their signs even
on extra sheets, right, if
196
00:13:01.529 --> 00:13:05.690
you can connect multiple day daters ways
ext and sheets. I mean it just
197
00:13:05.809 --> 00:13:09.289
depends on the volume at which you
are actually crunching this information, and you
198
00:13:09.370 --> 00:13:13.399
have services like a w as that
you can scale it up to enterprises or
199
00:13:13.440 --> 00:13:16.440
you can use for your small medium
businesses, not just state WS, Forsu,
200
00:13:16.720 --> 00:13:20.159
GCP, anything, anything in that
way. Yeah, I mean,
201
00:13:20.240 --> 00:13:22.519
long story shot, that's how they
kind of like transform businesses. I would
202
00:13:22.559 --> 00:13:28.230
call myself is kind of like a
distital transformation where you don't necessarily aware companies
203
00:13:28.230 --> 00:13:31.909
don't necessarily have still tools. We
transition them to district tools and everything.
204
00:13:31.990 --> 00:13:37.789
So I help them take confirmed decisions
to better make them customer experience and everything.
205
00:13:37.870 --> 00:13:39.350
So that's kind of like my day
to day job. But apart from
206
00:13:39.389 --> 00:13:43.779
that, there is the actually digita
sign stuff that we actually do as well,
207
00:13:43.899 --> 00:13:48.779
which is like building models or running
tests, like hypothesis testing and a
208
00:13:48.820 --> 00:13:52.100
lot of other bill learning test methods
that we actually do as well. That
209
00:13:52.220 --> 00:13:56.769
makes a lot of sense and actually
wanted to touch on like one of the
210
00:13:56.809 --> 00:14:01.690
last things that you said, the
transition from like the that these companies have
211
00:14:01.809 --> 00:14:05.850
to make you know, since ai
and aml a kind of in every part
212
00:14:05.970 --> 00:14:11.440
now, including like hr, accounting
everything right exactly. I wanted to ask
213
00:14:11.440 --> 00:14:16.200
you how hard is it to integrate
these data pipelines and companies that are seemingly
214
00:14:16.320 --> 00:14:20.840
resistant to wanting to adapt change,
since it does like messages with their entire
215
00:14:22.120 --> 00:14:26.190
like way of working? So what
trends have you noticed and, like should
216
00:14:26.190 --> 00:14:30.190
companies be more open to this like
going forward? Yeah, yeah, I
217
00:14:30.269 --> 00:14:33.309
think. I mean with how we
would affected in two thousand and twenty and
218
00:14:33.070 --> 00:14:37.549
the the need for you to kind
of like learn your customers. And you're
219
00:14:37.700 --> 00:14:41.340
not in just customers, but your
fit in the market. Right. You
220
00:14:41.379 --> 00:14:45.500
may have a lot of competition and
everything, but see, just holding your
221
00:14:45.539 --> 00:14:50.179
data in an Exu Shiet or somewhere
in a database and just taking your own
222
00:14:50.179 --> 00:14:52.450
informations never helps. I would ask
companies to kind of like think of it.
223
00:14:52.769 --> 00:14:56.690
Right. How are you taking decisions? Are they data driven or are
224
00:14:56.769 --> 00:15:01.649
they just experience driven? Right?
Or maybe you're on subject matter expert.
225
00:15:01.690 --> 00:15:05.610
You don't need to understand a lot
of your data, right, because multiple
226
00:15:05.649 --> 00:15:09.240
people, like, I mean there
are there are leaderships and there are also
227
00:15:09.279 --> 00:15:13.519
executive people who've been doing this for
the last forty years. You don't necessarily
228
00:15:13.559 --> 00:15:16.320
be doing the data sign stuff of
it, right, but in order for
229
00:15:16.399 --> 00:15:20.519
you to take a data driven decision, you probably may not need to spend
230
00:15:20.559 --> 00:15:26.149
a week generating a report of a
particular ask, but you can probably do
231
00:15:26.350 --> 00:15:30.710
that asap and like a few hours, right. So, in order for
232
00:15:30.830 --> 00:15:37.299
you to increase that operational efficiency and
on also maintain your team's operations efficiency,
233
00:15:37.379 --> 00:15:39.899
to kind of like learn every week, every day you can learn, test
234
00:15:39.940 --> 00:15:46.899
things out and take and the learning
outcomes that you take can be amplified by
235
00:15:46.059 --> 00:15:50.250
doing this on and on for like
another three thousand and sixteen nine day model.
236
00:15:50.490 --> 00:15:54.570
And you can make changes on your
on your applications, on your website,
237
00:15:54.690 --> 00:15:58.009
on how the customer facing contents as
well, so you to make better
238
00:15:58.090 --> 00:16:00.649
user experiences and everything. I strongly
believe. I think the transition comes in
239
00:16:00.730 --> 00:16:03.679
a way where you can work with
a POC, like a proof of concept.
240
00:16:03.799 --> 00:16:07.279
You take a subset of information and
take a take a business unit that
241
00:16:07.600 --> 00:16:12.559
needs to grow right and understand how
you are analyzing data, how are people
242
00:16:12.559 --> 00:16:17.720
analyzing their data right, and what
kind of volume that is coming with with
243
00:16:17.879 --> 00:16:22.029
a net to kind of like understand
the capacity at which you can take our
244
00:16:22.110 --> 00:16:26.710
informed decision and out of that you
can budget out and see if you need
245
00:16:26.870 --> 00:16:30.389
more resources. And there are companies
out there and that can help you get
246
00:16:30.470 --> 00:16:33.659
better at doing stuff. You don't
always have to kind of like have some
247
00:16:33.779 --> 00:16:37.940
many in house. You can have
consulting companies kind of like come to you
248
00:16:37.139 --> 00:16:41.220
and people like me, like independent
consultants, can come in like in a
249
00:16:41.299 --> 00:16:45.100
week or two and help you strategize
and build you a plan for you to
250
00:16:45.179 --> 00:16:51.690
execute right, and that will actually
help start off right. So that starting
251
00:16:51.730 --> 00:16:56.289
point is where companies have a lot
of inhibitions or they have some previous experiences
252
00:16:56.409 --> 00:17:00.129
that they don't necessarily want to because
obviously the companies do not necessarily have the
253
00:17:00.250 --> 00:17:06.480
time to implement something new. They
always have something very, very time taking
254
00:17:06.559 --> 00:17:08.759
out. It's probably very busy at
that time. So it's always good to
255
00:17:08.759 --> 00:17:11.839
kind of like have some expert come
in look at your from a from a
256
00:17:11.920 --> 00:17:15.390
high level, right. I call
it the EAGLIGHT. So someone someone from
257
00:17:15.390 --> 00:17:18.349
an from a high level, cause
you look at that and see the INS
258
00:17:18.390 --> 00:17:22.069
and out, see the gaps and
everything, and could give you a different
259
00:17:22.069 --> 00:17:26.150
perspective and that's how you take it
forward. I mean you will have to
260
00:17:26.309 --> 00:17:30.740
always evolve. You can just be
a traditional company always right. Yeah,
261
00:17:30.779 --> 00:17:33.059
absolutely, Dad. Thank you.
So I wanted to move on to talk
262
00:17:33.099 --> 00:17:37.500
about the data standard, and we
have the data standard here. Are Building
263
00:17:37.539 --> 00:17:41.180
a community of date enthusiasts and data
thought leaders. Everyone is open to collaboration
264
00:17:41.380 --> 00:17:45.410
and to connect with each other.
What can our community do for you?
265
00:17:45.490 --> 00:17:48.890
You guys are doing it done with
job. I mean this is obviously letting
266
00:17:48.130 --> 00:17:53.210
others talk. Shared experiences will probably
rule out a lot more simpler questions that
267
00:17:53.369 --> 00:17:57.599
usually students of people think it's really
hard. I think we're the the channel
268
00:17:57.640 --> 00:18:00.079
that you have, like a connect, a bridge between all of that,
269
00:18:00.240 --> 00:18:04.279
and we always wanted to kind of
like explore new opportunities as well. I
270
00:18:04.359 --> 00:18:07.920
mean, we'll be open to kind
of like discuss any new projects that you
271
00:18:08.039 --> 00:18:12.349
have, and I was always kind
of like mean competitions out there and there
272
00:18:12.390 --> 00:18:17.910
are also hacktons and everything that it's
really good for as a data scientist for
273
00:18:17.950 --> 00:18:22.190
you to kind of get that exposure, because that really helps in and if
274
00:18:22.190 --> 00:18:25.509
you work at an organization or if
you work at a startup or something like
275
00:18:25.589 --> 00:18:27.940
that, because you look at the
culture of the moment, data scientist is
276
00:18:29.059 --> 00:18:33.180
just not one feeling right and all
one kind of another thing that you do.
277
00:18:33.220 --> 00:18:36.420
You actually with many hats. You
you go to a marketing team,
278
00:18:36.460 --> 00:18:38.339
you actually we have to kind of
like get into their understanding, get into
279
00:18:38.380 --> 00:18:41.089
the line of understanding, to kind
of like get a gather their equinement.
280
00:18:41.170 --> 00:18:45.089
So it's not like you already have
like a Bible or template off questions that
281
00:18:45.170 --> 00:18:48.089
you ask every team and you get
the resources something. So I think that
282
00:18:48.210 --> 00:18:52.049
will probably also kind of like help
a lot in terms of like from a
283
00:18:52.130 --> 00:18:56.200
theoretical basis or a practical basis,
data standard. Students can also have this
284
00:18:56.720 --> 00:19:03.160
practical and real time experiences from people
and, yeah, and they'll probably help
285
00:19:03.200 --> 00:19:06.640
a lot and for them to transition
into better jobs. Oh Yeah, thank
286
00:19:06.680 --> 00:19:10.029
you for that and also thank you
for so much for coming on to our
287
00:19:10.069 --> 00:19:12.789
share today and we really really appreciate
your time and I enjoyed our conversation.
288
00:19:14.670 --> 00:19:17.630
But if we wanted to learn more
about you or talk to you more,
289
00:19:17.950 --> 00:19:21.750
where can our community find you?
Online? Online, you can actually scan
290
00:19:21.990 --> 00:19:25.460
the quo code. It's linked to
my website. Yeah, and you can
291
00:19:25.460 --> 00:19:29.220
find me on Linkedin. It's my
last name. You can find it on
292
00:19:29.259 --> 00:19:32.940
Linkedin as well, that's probably the
best way to find fame. We find
293
00:19:33.059 --> 00:19:36.700
find out and if people listening to
the podcast as well, if you have
294
00:19:36.740 --> 00:19:40.490
any questions or if you have any
need Genera the advisor suggestion, as you
295
00:19:40.529 --> 00:19:42.890
can always go on to my website
or you're looking, bring me on Linkedin
296
00:19:42.970 --> 00:19:47.049
and we can actually have a chat
in this this is an open field at
297
00:19:47.089 --> 00:19:52.089
the moment and it's a always help. I'm always like to share my experiences
298
00:19:52.210 --> 00:19:55.279
and learn from you guys as well. Hey, thank you so much for
299
00:19:55.440 --> 00:19:59.079
information and for more information about the
data standard, you can find us at
300
00:19:59.160 --> 00:20:03.240
www dot data standard. Thank you
so much for joining us and I hope
301
00:20:03.279 --> 00:20:04.359
to catch you next time. Thank
you, braying. Good bye.
1
00:00:10.269 --> 00:00:13.470
Hi, I'm your hus premier'll be
sunker, and this is the data standard
2
00:00:13.470 --> 00:00:17.030
audio experience. We have a very
exciting guest today on our podcast. We
3
00:00:17.109 --> 00:00:21.859
have anthony a Kassatti, data scientists
and global search strategies. Today we're speaking
4
00:00:21.859 --> 00:00:26.780
about his journey into data science,
as well as applications within small beating manufacturing.
5
00:00:27.140 --> 00:00:29.300
Welcome to the show. Thank you. Thank you for having me.
6
00:00:29.579 --> 00:00:32.619
Yeah, of course, thanks for
coming on to the show today. Go
7
00:00:32.740 --> 00:00:35.969
ahead and introduce yourself. Sure so, like I said in, my name
8
00:00:36.009 --> 00:00:39.530
is anthony a Kastadi. I am
an independent data scientist and I work with
9
00:00:39.929 --> 00:00:45.450
a lot of marketing teams and data
science teams at various manufacturers or technology companies
10
00:00:45.490 --> 00:00:51.560
and healthcare destroy myself to create a
new methodologies and technologies to help them kind
11
00:00:51.600 --> 00:00:55.600
of like take a step further in
the end others. Yeah, definitely.
12
00:00:55.640 --> 00:00:59.719
Well, I'm really excited to have
you for this conversation today, and your
13
00:00:59.799 --> 00:01:03.439
career thus far has been incredibly impressive, with you first starting as an electrical
14
00:01:03.520 --> 00:01:07.709
engineer and then making the transition to
like an independent data science consultant, as
15
00:01:07.750 --> 00:01:10.790
you are now. How did you
end up working in data science and what
16
00:01:10.909 --> 00:01:14.189
has been the hardest part of transition. So yeah, I mean I mean
17
00:01:14.629 --> 00:01:18.269
for the past ten to eleven years
it never occurred to me that I'll be
18
00:01:18.469 --> 00:01:22.579
here like as a data scientist,
and it was not even a job back
19
00:01:22.620 --> 00:01:26.180
then, right. So when I
was finishing my engineering I had to look
20
00:01:26.180 --> 00:01:30.620
for work and at that time there
was a boom in terms of like searching
21
00:01:30.739 --> 00:01:36.569
and optimization and they were people selling
services online and kind of like. That
22
00:01:36.849 --> 00:01:42.489
increased the visibility for web developers and
software engineers to build new applications and everything.
23
00:01:42.530 --> 00:01:47.010
So that's how I went entered into
the web development field. So I
24
00:01:47.090 --> 00:01:49.120
worked at a marketing agency in Melbourne. Could the good thing is and in
25
00:01:49.239 --> 00:01:53.560
engineering you actually are you learn.
It's a multi faceted course. You you
26
00:01:53.640 --> 00:01:57.480
learn computer science, you'll learn engineering, you also learn a little bit of
27
00:01:57.599 --> 00:02:01.790
business administration and everything in there.
So the this part of me always like
28
00:02:01.989 --> 00:02:07.189
programming, mathematics and everything. So
as a developer I was able to kind
29
00:02:07.189 --> 00:02:09.550
of like get into like the data
collection process. I was able to build
30
00:02:09.550 --> 00:02:14.870
multiple software applications and everything and slowly
over the course of ten years, I
31
00:02:14.949 --> 00:02:19.219
mean you you ended up being a
data signs them. Yeah, it's a
32
00:02:19.300 --> 00:02:22.860
cool it's a cool thing to have, like your data signers, but it's
33
00:02:22.860 --> 00:02:24.580
also like a cool job as well
then. Yeah, yeah, definitely.
34
00:02:24.979 --> 00:02:30.129
And what's been like, at least
for you? The hardest part about being
35
00:02:30.169 --> 00:02:35.370
a data scientists? I mean the
hardest part is it is so to keep
36
00:02:35.370 --> 00:02:39.770
it to keep it simple. You
there are multiple different segments within the data
37
00:02:39.810 --> 00:02:44.090
signs, right. So when you
branch out a lip, step back a
38
00:02:44.090 --> 00:02:46.159
little bit, you look at Ai
am all in deep learning, and you
39
00:02:46.680 --> 00:02:53.719
the software engineers, database administrators or
people who are using people who are kind
40
00:02:53.719 --> 00:02:57.000
of like doing the data mine stuff. Now it broken down into like architects,
41
00:02:57.080 --> 00:03:00.310
engineers and data signed this kind of
thing. So the hardest part is
42
00:03:00.349 --> 00:03:02.229
to kind of like you find your
fit, right, to find your fit
43
00:03:02.310 --> 00:03:07.509
in an organization. So, if
you're trying to transition, do you want
44
00:03:07.509 --> 00:03:10.030
to be data scientists? Do you
want it to I mean say if you
45
00:03:10.110 --> 00:03:14.379
want to, if you're an engineer
or an analyst, that there are certain
46
00:03:14.460 --> 00:03:17.219
things that you always do want a
day to be basis right and you don't
47
00:03:17.259 --> 00:03:21.780
want to kind of like be there
for a longer time because they are to
48
00:03:21.860 --> 00:03:24.219
science. Is We're growing. I
mean in the last five years itself,
49
00:03:24.860 --> 00:03:30.250
we've outgrown ourselves in terms of writing
algorithms and everything, and basically we are
50
00:03:30.409 --> 00:03:35.569
making ourselves out of business right,
because that you now have. You not
51
00:03:35.729 --> 00:03:40.719
have the tools and libraries where you
can run thousands of algorithms and ten minutes.
52
00:03:42.120 --> 00:03:45.039
Back in the day, I mean
like Justin like last three years,
53
00:03:45.159 --> 00:03:47.360
I actually took probably a day or
two to kind of like write an algorithm.
54
00:03:47.479 --> 00:03:52.319
Right. Normally arena regulation not came. Came means clustering or something like
55
00:03:52.360 --> 00:03:53.960
that. Now it's like you have
ten different things that you can do at
56
00:03:53.960 --> 00:03:57.509
the same time. So I think
the hardest part is to wear is to
57
00:03:57.590 --> 00:03:59.669
kind of like you find your fit. What do you want to kind of
58
00:03:59.710 --> 00:04:01.389
like move forward? And where are
you doing that? I mean you are
59
00:04:01.389 --> 00:04:04.069
you doing at a a small scale, or you're doing at a bigger scale?
60
00:04:04.069 --> 00:04:10.379
Are you working at the Amazons,
the Google kind of like space,
61
00:04:10.460 --> 00:04:14.979
or are you actually trying to kind
of like improve someone's business operations, improve
62
00:04:15.060 --> 00:04:17.579
customer experiences or something like that?
So that's kind of like the way thought
63
00:04:17.620 --> 00:04:20.379
process that you need to have.
Huh, that's interesting. So like finding
64
00:04:20.420 --> 00:04:25.490
your niche is really important within data
science. Yeah, every day that,
65
00:04:25.529 --> 00:04:29.490
I mean not all day is very
common, and even even the volume and
66
00:04:29.529 --> 00:04:32.209
the wallacity at which the data comes
in is also not common. So you
67
00:04:32.610 --> 00:04:36.970
never can actually work at a small
scale company, go directly to like an
68
00:04:36.970 --> 00:04:43.079
Amazon kind of service and figure out
that I can. So you're always bombarded
69
00:04:43.160 --> 00:04:46.120
with something very new and something really
exciting. Always yeah, well, that's
70
00:04:46.160 --> 00:04:51.040
really cool. And do you think
coming from that background helped you anyway,
71
00:04:51.319 --> 00:04:57.509
or it was it just like what
you learned within the within your way as
72
00:04:57.550 --> 00:05:00.829
a becoming a data scientist? I
me yes, it absolutely does. I
73
00:05:00.949 --> 00:05:06.699
mean having an engineering background absolutely helps
in transitioning to the data signs we are
74
00:05:08.379 --> 00:05:13.019
but I would not say I would
probably recommend it, because I have friends
75
00:05:13.060 --> 00:05:16.300
and family and also colleagues who come
from different verticals, who come from accounting.
76
00:05:16.699 --> 00:05:20.569
Our data analysts, right, who
are who done the hut on psychology,
77
00:05:20.649 --> 00:05:24.810
and they are also kind of like
there aren't analyzing customer behavior, and
78
00:05:24.850 --> 00:05:28.410
also they are, there are a
type of data assigned to themselves. I
79
00:05:28.529 --> 00:05:30.889
mean, if you're understanding the code
of data science, you need to always
80
00:05:30.970 --> 00:05:36.319
have an interest in math rate because
you're more of a statistician than software engineer.
81
00:05:36.399 --> 00:05:41.199
You're better right, if you're better
like that. So you think of
82
00:05:41.279 --> 00:05:44.000
it in that way and depending on
the job as well, you if you
83
00:05:44.240 --> 00:05:46.079
look at the job descriptions or the
companies to kind of like hire you.
84
00:05:46.279 --> 00:05:50.269
Sometimes they ask you for like an
engineering but some who you ask you for
85
00:05:50.509 --> 00:05:54.829
like a PhD in statistics of instead
physics or something like that. So it's
86
00:05:54.829 --> 00:05:58.509
always about the job. where,
where are you going in and where are
87
00:05:58.550 --> 00:06:01.269
you kind of like placing our so
and engineering background always helps. I mean
88
00:06:01.310 --> 00:06:06.980
that's how it helped me transition from
a completely filled job like an engineering like
89
00:06:08.100 --> 00:06:10.699
you go out on the field,
get the job done and everything. But
90
00:06:10.939 --> 00:06:15.060
for me it eventually turned out that
I like analytics, I like rip,
91
00:06:15.459 --> 00:06:17.850
I like how numbers right. I
have to make narratives out of it,
92
00:06:18.009 --> 00:06:20.569
meaning out of it, get tell
story out of it. So started to
93
00:06:20.610 --> 00:06:25.170
help businesses, started to see the
change. Obviously, as an engineer,
94
00:06:25.250 --> 00:06:28.250
I wanted to do something and I
wanted to see the effects of it so
95
00:06:28.449 --> 00:06:31.290
I can have feel pleasant right,
I mean there's something in the world that
96
00:06:31.329 --> 00:06:35.160
I'm actually doing, that it's actually
working. So it happens in data signs,
97
00:06:35.160 --> 00:06:40.279
I mean you can now that everything
is instant in real time, with
98
00:06:40.600 --> 00:06:44.079
a lot of data that's coming in. I mean you don't necessarily have any
99
00:06:44.199 --> 00:06:47.269
filters everywhere. So it is a
truly good experience that you can have with
100
00:06:47.430 --> 00:06:51.470
this one. I mean, even
though you compaplate. I mean your story
101
00:06:51.550 --> 00:06:55.709
is really cool because a lot of
our listeners are students that are trying to
102
00:06:55.750 --> 00:06:59.470
break into data scientists, or at
least people that are within different industries that
103
00:06:59.509 --> 00:07:01.500
are trying to, you know,
get into data science because they see that
104
00:07:01.579 --> 00:07:05.660
it's a like a really cool industry. So for those people, do you
105
00:07:05.819 --> 00:07:11.139
have any advice or anything that,
like when you first came to the industry,
106
00:07:11.139 --> 00:07:14.500
where there any misconceptions that she had
to like get over that yourself,
107
00:07:14.740 --> 00:07:16.930
or any advice that you'd give to
people such as that trying to break into
108
00:07:16.930 --> 00:07:20.370
the industry? Yeah, I think
that in my early days, we're a
109
00:07:21.009 --> 00:07:25.370
C shop. Java had do big
data. This was kind of like a
110
00:07:25.930 --> 00:07:30.959
big, big thing, and so
you start to see companies, recruiters and
111
00:07:30.600 --> 00:07:33.480
people who hire you kind of like
ask you for everything, right, I
112
00:07:33.879 --> 00:07:36.079
you could a job, or you
could at C plus ps, I could
113
00:07:36.120 --> 00:07:40.480
at Python, I mean anything that
comes in right. So the misconception is
114
00:07:40.800 --> 00:07:44.279
be good at something, don't just
for the sake of it, if there's
115
00:07:44.279 --> 00:07:46.069
something coming in the industry or if
they if there's a hype about it,
116
00:07:46.230 --> 00:07:49.389
if there's a new programming language that
you have to learn for it, and
117
00:07:49.589 --> 00:07:53.949
probably Yah who's written something, or
a Google as as brought in some code.
118
00:07:54.350 --> 00:07:56.910
You don't necessarily have to kind of
like jump in right away about it,
119
00:07:56.990 --> 00:08:00.259
but just do your research on it, look at how that affects your
120
00:08:00.459 --> 00:08:03.420
current job and look at your career
progression as well, kind of like go
121
00:08:03.660 --> 00:08:07.980
through the forums and everything. But
the way that I was trying, I
122
00:08:07.060 --> 00:08:11.180
mean I was miss come misconceptualized,
is I have to learn everything after be
123
00:08:11.259 --> 00:08:13.649
good at good, be at good, be good at everything so I can
124
00:08:13.970 --> 00:08:16.769
take the interviews properly, get to
get the job done and everything like that.
125
00:08:16.970 --> 00:08:20.649
But it's always it's the company's choice
to kind of like hire you or
126
00:08:20.689 --> 00:08:24.089
not. Right. You may be
good at our or you may be good
127
00:08:24.129 --> 00:08:26.970
at the Python, but maybe the
company can come and ask I only need
128
00:08:26.089 --> 00:08:28.800
people for Jova. That doesn't make
Oud. You actually have to be like
129
00:08:30.000 --> 00:08:33.879
hundred percent expert in Java like that. So think about it, whether it's
130
00:08:33.879 --> 00:08:37.279
a right choice or not, and
they don't owe stress yourself and data sense.
131
00:08:37.360 --> 00:08:39.799
Is Not a job to get stressed. It's a cool job. You
132
00:08:39.919 --> 00:08:43.669
have to keep your brain pleasant because
you'll have to keep thinking always, because
133
00:08:43.990 --> 00:08:48.549
you have people come to you with
their requirements, with their problems. You
134
00:08:48.789 --> 00:08:50.750
probably need to find a solution with
those numbers. Rate, you are the
135
00:08:50.830 --> 00:08:56.220
only person that they rely on to
get the expert feedback one right. So
136
00:08:56.500 --> 00:08:58.820
it's not always that you have a
magic had and you take a rye bit
137
00:08:58.899 --> 00:09:01.700
out of it, but it's always
kind of like you have to kind of
138
00:09:01.740 --> 00:09:05.460
like make sense out of these numbers, and that's why they rely on you.
139
00:09:05.700 --> 00:09:09.259
Right. So I would say that's
kind of like the first advice I
140
00:09:09.299 --> 00:09:11.090
would like, the first in the
last eddy I would give you that will
141
00:09:11.169 --> 00:09:16.730
actually shape you to being a better
person at me taking making decisions and taking
142
00:09:16.809 --> 00:09:20.090
decisions. That will also give you
a effect in terms of how you work
143
00:09:20.210 --> 00:09:24.210
in your workplace as well. Hey, yeah, thank you. That makes
144
00:09:24.360 --> 00:09:28.080
a lot of sense. I actually
wanted to switch over from like your pastory
145
00:09:28.240 --> 00:09:31.799
to like the work that you're currently
doing. I understand that you mostly deal
146
00:09:31.879 --> 00:09:35.440
with small menium manufactures and, as
someone who doesn't really know too much about
147
00:09:35.440 --> 00:09:39.629
the industry, would you mind explaining
the biggest challenges that you face and also,
148
00:09:39.909 --> 00:09:43.549
like the word that you do on
the clients that you work with.
149
00:09:43.950 --> 00:09:46.549
Yeah, sure, sure, yes, small menuon manufacturers is one of the
150
00:09:46.590 --> 00:09:50.590
niche that I ply worked in the
experience that I had in the US,
151
00:09:50.789 --> 00:09:54.379
and I also worked with less than
fifty million dollar raven new companies in here
152
00:09:54.539 --> 00:10:00.059
are and I also work with smaller
companies as well, right, I mean
153
00:10:00.100 --> 00:10:03.419
these are companies that are attacked to
consumer. You to see kind of department
154
00:10:05.179 --> 00:10:07.289
like a background. Out of all
of this, I figured one thing that
155
00:10:07.450 --> 00:10:11.090
is data everywhere that the company necessarily
kind of like consumes, right, so
156
00:10:11.610 --> 00:10:15.450
the type of their time and how
they consume is not necessarily thought of.
157
00:10:16.169 --> 00:10:20.049
So you just have lots and lots
of volume coming in, but you never
158
00:10:20.169 --> 00:10:26.519
necessarily understood how to kind of like
strategize structure it for your consumption and how
159
00:10:26.559 --> 00:10:28.679
do you want it to take up
learn, like the build, test and
160
00:10:28.799 --> 00:10:33.480
learning model is prop is not properly
thought of. So what I've done basically
161
00:10:33.639 --> 00:10:37.350
was to kind of like come in
and you have a lot. You have
162
00:10:37.389 --> 00:10:39.429
a lot of content out there for
data science. You have a lot of
163
00:10:39.470 --> 00:10:41.830
youtube videos, you a lot of
people talking about it. You have a
164
00:10:41.870 --> 00:10:46.590
lot of food cams and exercises that
people can take to learn theory tically what
165
00:10:46.789 --> 00:10:48.909
their assience is. And then you, if you wanted to kind of implement
166
00:10:48.950 --> 00:10:54.860
that in your workplace. You probably
may not necessarily take a cost effective cost
167
00:10:54.940 --> 00:11:00.019
division to decision. Right, you
probably are choosing for tools that are beyond
168
00:11:00.139 --> 00:11:05.250
what you need. Right. Yeah, it's just like you using Microsoft word,
169
00:11:05.649 --> 00:11:07.970
but you only use like a few
tools, and microsoftware, you don't
170
00:11:07.970 --> 00:11:11.649
use its whole spectrum of things.
They're right. So, so what I've
171
00:11:11.690 --> 00:11:16.129
done basically is to kind of like
it's if you go in, you understand
172
00:11:16.129 --> 00:11:20.639
the business needs, right. So
most of these companies they have transactional data,
173
00:11:20.720 --> 00:11:24.440
they have customer behavior data, customer
care data that coming from their phone
174
00:11:24.480 --> 00:11:28.600
calls or the contact form their inquiries. Some has warranty issues, some may
175
00:11:28.679 --> 00:11:33.350
have product problems, right, so
all of this has to be mind in
176
00:11:33.389 --> 00:11:37.070
terms of understanding to innovate your product. Basically, what would be my next
177
00:11:37.149 --> 00:11:43.389
iteration rate? And that's where I
come in when I set up master data
178
00:11:43.429 --> 00:11:50.100
management pipelines where we create ways in
which we channelize these fountains of information into
179
00:11:50.259 --> 00:11:54.899
silos, right, and depending on
different teams you may have anop you and
180
00:11:54.100 --> 00:11:56.820
operations team, you have Ay,
you may have an e commerce team,
181
00:11:56.820 --> 00:12:01.889
you have have pricing teams, you
have engineering teams that you always have to
182
00:12:01.970 --> 00:12:07.210
feed information, right. That's how
they take informed decisions and everyone is a
183
00:12:07.330 --> 00:12:09.409
part of the state ASIGENCE team.
It's not like dat designers is only kind
184
00:12:09.409 --> 00:12:13.850
of liquing everything, right. So
I kind of like bring that culture in
185
00:12:15.330 --> 00:12:20.879
as well and for company he's who
never had these experiences. Obviously a shift
186
00:12:20.960 --> 00:12:26.879
right in an agile way of doing
things. So it's obviously a practice that
187
00:12:26.919 --> 00:12:30.710
you'll have to keep doing it and
show the outcome of it. And and
188
00:12:31.549 --> 00:12:35.230
companies are actually going trying to take
that transition as well. So you don't
189
00:12:35.230 --> 00:12:39.070
need to invest heavily on data signers
event there's a misconstruption, a misconception within
190
00:12:39.149 --> 00:12:43.190
the organizations where you need get aware
of houses. You need etl tools,
191
00:12:43.269 --> 00:12:48.019
you probably need tools and tool sets
and services. You any licenses and and
192
00:12:48.500 --> 00:12:52.460
it's an anyways. It's an up
upwards to a million to like couple of
193
00:12:52.539 --> 00:12:56.539
million dollars a year in terms of
like the budget that the company makes.
194
00:12:56.580 --> 00:12:58.610
But I don't necessarily feel that is
important. I mean, at the moment,
195
00:12:58.730 --> 00:13:01.529
you can actually do their signs even
on extra sheets, right, if
196
00:13:01.529 --> 00:13:05.690
you can connect multiple day daters ways
ext and sheets. I mean it just
197
00:13:05.809 --> 00:13:09.289
depends on the volume at which you
are actually crunching this information, and you
198
00:13:09.370 --> 00:13:13.399
have services like a w as that
you can scale it up to enterprises or
199
00:13:13.440 --> 00:13:16.440
you can use for your small medium
businesses, not just state WS, Forsu,
200
00:13:16.720 --> 00:13:20.159
GCP, anything, anything in that
way. Yeah, I mean,
201
00:13:20.240 --> 00:13:22.519
long story shot, that's how they
kind of like transform businesses. I would
202
00:13:22.559 --> 00:13:28.230
call myself is kind of like a
distital transformation where you don't necessarily aware companies
203
00:13:28.230 --> 00:13:31.909
don't necessarily have still tools. We
transition them to district tools and everything.
204
00:13:31.990 --> 00:13:37.789
So I help them take confirmed decisions
to better make them customer experience and everything.
205
00:13:37.870 --> 00:13:39.350
So that's kind of like my day
to day job. But apart from
206
00:13:39.389 --> 00:13:43.779
that, there is the actually digita
sign stuff that we actually do as well,
207
00:13:43.899 --> 00:13:48.779
which is like building models or running
tests, like hypothesis testing and a
208
00:13:48.820 --> 00:13:52.100
lot of other bill learning test methods
that we actually do as well. That
209
00:13:52.220 --> 00:13:56.769
makes a lot of sense and actually
wanted to touch on like one of the
210
00:13:56.809 --> 00:14:01.690
last things that you said, the
transition from like the that these companies have
211
00:14:01.809 --> 00:14:05.850
to make you know, since ai
and aml a kind of in every part
212
00:14:05.970 --> 00:14:11.440
now, including like hr, accounting
everything right exactly. I wanted to ask
213
00:14:11.440 --> 00:14:16.200
you how hard is it to integrate
these data pipelines and companies that are seemingly
214
00:14:16.320 --> 00:14:20.840
resistant to wanting to adapt change,
since it does like messages with their entire
215
00:14:22.120 --> 00:14:26.190
like way of working? So what
trends have you noticed and, like should
216
00:14:26.190 --> 00:14:30.190
companies be more open to this like
going forward? Yeah, yeah, I
217
00:14:30.269 --> 00:14:33.309
think. I mean with how we
would affected in two thousand and twenty and
218
00:14:33.070 --> 00:14:37.549
the the need for you to kind
of like learn your customers. And you're
219
00:14:37.700 --> 00:14:41.340
not in just customers, but your
fit in the market. Right. You
220
00:14:41.379 --> 00:14:45.500
may have a lot of competition and
everything, but see, just holding your
221
00:14:45.539 --> 00:14:50.179
data in an Exu Shiet or somewhere
in a database and just taking your own
222
00:14:50.179 --> 00:14:52.450
informations never helps. I would ask
companies to kind of like think of it.
223
00:14:52.769 --> 00:14:56.690
Right. How are you taking decisions? Are they data driven or are
224
00:14:56.769 --> 00:15:01.649
they just experience driven? Right?
Or maybe you're on subject matter expert.
225
00:15:01.690 --> 00:15:05.610
You don't need to understand a lot
of your data, right, because multiple
226
00:15:05.649 --> 00:15:09.240
people, like, I mean there
are there are leaderships and there are also
227
00:15:09.279 --> 00:15:13.519
executive people who've been doing this for
the last forty years. You don't necessarily
228
00:15:13.559 --> 00:15:16.320
be doing the data sign stuff of
it, right, but in order for
229
00:15:16.399 --> 00:15:20.519
you to take a data driven decision, you probably may not need to spend
230
00:15:20.559 --> 00:15:26.149
a week generating a report of a
particular ask, but you can probably do
231
00:15:26.350 --> 00:15:30.710
that asap and like a few hours, right. So, in order for
232
00:15:30.830 --> 00:15:37.299
you to increase that operational efficiency and
on also maintain your team's operations efficiency,
233
00:15:37.379 --> 00:15:39.899
to kind of like learn every week, every day you can learn, test
234
00:15:39.940 --> 00:15:46.899
things out and take and the learning
outcomes that you take can be amplified by
235
00:15:46.059 --> 00:15:50.250
doing this on and on for like
another three thousand and sixteen nine day model.
236
00:15:50.490 --> 00:15:54.570
And you can make changes on your
on your applications, on your website,
237
00:15:54.690 --> 00:15:58.009
on how the customer facing contents as
well, so you to make better
238
00:15:58.090 --> 00:16:00.649
user experiences and everything. I strongly
believe. I think the transition comes in
239
00:16:00.730 --> 00:16:03.679
a way where you can work with
a POC, like a proof of concept.
240
00:16:03.799 --> 00:16:07.279
You take a subset of information and
take a take a business unit that
241
00:16:07.600 --> 00:16:12.559
needs to grow right and understand how
you are analyzing data, how are people
242
00:16:12.559 --> 00:16:17.720
analyzing their data right, and what
kind of volume that is coming with with
243
00:16:17.879 --> 00:16:22.029
a net to kind of like understand
the capacity at which you can take our
244
00:16:22.110 --> 00:16:26.710
informed decision and out of that you
can budget out and see if you need
245
00:16:26.870 --> 00:16:30.389
more resources. And there are companies
out there and that can help you get
246
00:16:30.470 --> 00:16:33.659
better at doing stuff. You don't
always have to kind of like have some
247
00:16:33.779 --> 00:16:37.940
many in house. You can have
consulting companies kind of like come to you
248
00:16:37.139 --> 00:16:41.220
and people like me, like independent
consultants, can come in like in a
249
00:16:41.299 --> 00:16:45.100
week or two and help you strategize
and build you a plan for you to
250
00:16:45.179 --> 00:16:51.690
execute right, and that will actually
help start off right. So that starting
251
00:16:51.730 --> 00:16:56.289
point is where companies have a lot
of inhibitions or they have some previous experiences
252
00:16:56.409 --> 00:17:00.129
that they don't necessarily want to because
obviously the companies do not necessarily have the
253
00:17:00.250 --> 00:17:06.480
time to implement something new. They
always have something very, very time taking
254
00:17:06.559 --> 00:17:08.759
out. It's probably very busy at
that time. So it's always good to
255
00:17:08.759 --> 00:17:11.839
kind of like have some expert come
in look at your from a from a
256
00:17:11.920 --> 00:17:15.390
high level, right. I call
it the EAGLIGHT. So someone someone from
257
00:17:15.390 --> 00:17:18.349
an from a high level, cause
you look at that and see the INS
258
00:17:18.390 --> 00:17:22.069
and out, see the gaps and
everything, and could give you a different
259
00:17:22.069 --> 00:17:26.150
perspective and that's how you take it
forward. I mean you will have to
260
00:17:26.309 --> 00:17:30.740
always evolve. You can just be
a traditional company always right. Yeah,
261
00:17:30.779 --> 00:17:33.059
absolutely, Dad. Thank you.
So I wanted to move on to talk
262
00:17:33.099 --> 00:17:37.500
about the data standard, and we
have the data standard here. Are Building
263
00:17:37.539 --> 00:17:41.180
a community of date enthusiasts and data
thought leaders. Everyone is open to collaboration
264
00:17:41.380 --> 00:17:45.410
and to connect with each other.
What can our community do for you?
265
00:17:45.490 --> 00:17:48.890
You guys are doing it done with
job. I mean this is obviously letting
266
00:17:48.130 --> 00:17:53.210
others talk. Shared experiences will probably
rule out a lot more simpler questions that
267
00:17:53.369 --> 00:17:57.599
usually students of people think it's really
hard. I think we're the the channel
268
00:17:57.640 --> 00:18:00.079
that you have, like a connect, a bridge between all of that,
269
00:18:00.240 --> 00:18:04.279
and we always wanted to kind of
like explore new opportunities as well. I
270
00:18:04.359 --> 00:18:07.920
mean, we'll be open to kind
of like discuss any new projects that you
271
00:18:08.039 --> 00:18:12.349
have, and I was always kind
of like mean competitions out there and there
272
00:18:12.390 --> 00:18:17.910
are also hacktons and everything that it's
really good for as a data scientist for
273
00:18:17.950 --> 00:18:22.190
you to kind of get that exposure, because that really helps in and if
274
00:18:22.190 --> 00:18:25.509
you work at an organization or if
you work at a startup or something like
275
00:18:25.589 --> 00:18:27.940
that, because you look at the
culture of the moment, data scientist is
276
00:18:29.059 --> 00:18:33.180
just not one feeling right and all
one kind of another thing that you do.
277
00:18:33.220 --> 00:18:36.420
You actually with many hats. You
you go to a marketing team,
278
00:18:36.460 --> 00:18:38.339
you actually we have to kind of
like get into their understanding, get into
279
00:18:38.380 --> 00:18:41.089
the line of understanding, to kind
of like get a gather their equinement.
280
00:18:41.170 --> 00:18:45.089
So it's not like you already have
like a Bible or template off questions that
281
00:18:45.170 --> 00:18:48.089
you ask every team and you get
the resources something. So I think that
282
00:18:48.210 --> 00:18:52.049
will probably also kind of like help
a lot in terms of like from a
283
00:18:52.130 --> 00:18:56.200
theoretical basis or a practical basis,
data standard. Students can also have this
284
00:18:56.720 --> 00:19:03.160
practical and real time experiences from people
and, yeah, and they'll probably help
285
00:19:03.200 --> 00:19:06.640
a lot and for them to transition
into better jobs. Oh Yeah, thank
286
00:19:06.680 --> 00:19:10.029
you for that and also thank you
for so much for coming on to our
287
00:19:10.069 --> 00:19:12.789
share today and we really really appreciate
your time and I enjoyed our conversation.
288
00:19:14.670 --> 00:19:17.630
But if we wanted to learn more
about you or talk to you more,
289
00:19:17.950 --> 00:19:21.750
where can our community find you?
Online? Online, you can actually scan
290
00:19:21.990 --> 00:19:25.460
the quo code. It's linked to
my website. Yeah, and you can
291
00:19:25.460 --> 00:19:29.220
find me on Linkedin. It's my
last name. You can find it on
292
00:19:29.259 --> 00:19:32.940
Linkedin as well, that's probably the
best way to find fame. We find
293
00:19:33.059 --> 00:19:36.700
find out and if people listening to
the podcast as well, if you have
294
00:19:36.740 --> 00:19:40.490
any questions or if you have any
need Genera the advisor suggestion, as you
295
00:19:40.529 --> 00:19:42.890
can always go on to my website
or you're looking, bring me on Linkedin
296
00:19:42.970 --> 00:19:47.049
and we can actually have a chat
in this this is an open field at
297
00:19:47.089 --> 00:19:52.089
the moment and it's a always help. I'm always like to share my experiences
298
00:19:52.210 --> 00:19:55.279
and learn from you guys as well. Hey, thank you so much for
299
00:19:55.440 --> 00:19:59.079
information and for more information about the
data standard, you can find us at
300
00:19:59.160 --> 00:20:03.240
www dot data standard. Thank you
so much for joining us and I hope
301
00:20:03.279 --> 00:20:04.359
to catch you next time. Thank
you, braying. Good bye.
