Friday 26 October 2018

TOP 10 THINGS YOU SHOULD AWARE OF BEFORE SHIFTING YOUR CAREER TO DATA SCIENCE/MACHINE LEARNING/AI


Hi Folks ,

As you are here I could understand that you are pretty much enthusiast to make your career transact (start) into data science / Machine learning.
It is very normal that so many questions and confusions you already have.
such as::
What is exactly ML(machine learning )? Is it same as data science? Do I am capable for this?
What should I learn more? What extra effort I need to give?

Relax, At the end of this article you will have a better vision towards present booming technology and the sexiest job career. 

Before moving directly to article, just to note that, I have subdivided these crucial 10 points into two categories as per our initial (pre-mature) thought process.

Logical aspects:

Before deciding a career, switch or start you should ask these points as a question to you and rate yourself where you are. The score is your ultimate milestone.

1.      Present career:

You may be involved to any business, environments or technologies but what it matters first that ,  are you presently on such a platform where you can transit yourself ?
The answer is “Yes” transaction is possible if you are only on relevant sectors having a decent technical knowledge/skill, the most preferable is having the background of CS/IT or a typical software engineer having experienced on real world applications (doesn’t matters how experienced you are). 

** career into IT/ITES or background of CS/IT


2.      Long term career goal:

Yes, like every other sector, a long term career goal is an important aspects here too. why?
People often thought of switch from technologies to business or vice versa whenever they couldn’t find expect-able growth on their career. But here if you are a real machine learning enthusiast and experienced the ultimate goal is “to transform yourself into a “data scientist””. So, if you have any long-term plans to be a business manager or vice president of any XYZ company so this might not help you on that. But trends are changing now a day. Most of the crucial companies want their mangers to be well skilled and trained with cutting edge technologies.

**don’t hesitate to code after having 15 years of experience 


3.      What technologies you are aware of

Just as simple as it is “The more you know the better it helps”. The exact programming skills you need is python/R (what data scientist prefer). But the other skills such as visualization, Cloud, API, web framework or any short of other programming language will help you much on understanding the AI application in your way and grabbing the various ideas to make your approaches better and polished. So, if you already worked on these before it’s an added advance for you if not the basic (Python/R) you need to add into your bucket. (don’t worry python is really easy as compared to other languages)


**must learn Python / R

4.      Willing to learn every odd day?

Learning and working have its own different flavor. Whenever you will work on a real time data or real environment you will face tons of huddle and challenges and remember every instance will let you learn a lesson. So it is very often that you need to be open for learn new things and the changing world. Just ask the same question am I ready for it ?

**Eager to learn new things.

5.      How open for coding and logical thinking?

It is most seen in Indian IT industries the willingness to code is gradually decreases with the increase of experience. which ultimately affects to your logical thinking and dependability.
A Big NO for that thought. You should always set your flag green to any short of development with direct involvement in root level coding and architecture which makes you understand the flow and process complete in and out.

**No hesitation for coading. 

6.      How strong you are in high school math ( kya app 10wi class se tez ho )

Probably you all have gone through aptitude and reasoning test where you may find the 10th level math. Do we really need that too here ?
No you don’t need that exactly but you need the base which covers the topic such as probabilities, statistics, derivatives, logarithms etc. .It’s not that tough, it just need a brush up, you can start at any of time whenever you think  of it.

** Concepts let you understand the algorithm deeply and helps on data pre-processing. (which is a part of model building) 



Technical aspects:

7.      Python/R, SQL:

As I mentioned earlier you must learn either or python or R which helps you building various models (applying the algorithms to complex structured and unstructured data).
Now the question is am I skilled enough to learn Python?
The answer is “Yes” if you already worked on or having knowledge on any programming language. I would say python is the easiest language and having largest community with numerous built in libraries at present date, the only thing is you just need to refresh your concepts in a new way.
Again question is there why R?
R is also a programming language like python but specially developed for data science .It also have a larger community and free accessible built in libraries but the only major difference of python and R is integrations with real time applications . Where Python is more reliable and stable as compared to R.

Here is few analysis by KDnuggets poll (You can google the difference of python and R )

Okay now why SQL??
It’s not about only SQL, the most important thing before going for any short of machine learning model or algorithm is the data and in real world most of the structured data are fetched through SQL queries. Yes you should have at least basic knowledge of any short of database at your initial stage .

**hands on experience on python/R along with database.

8.      Knowledge on visualization tools (Qlik view / Tablue ) :

There is no meaning if you have bunch of data and we don’t know how to represent it. Hence representation is the vital part of any kind of analysis before going for ML algorithms for any short of outcome and those representation also knows as visualization, which are driven by two basic and important tools such as Qlik view and Tablue.
Although python, R, SAAS (tool equivalent as R ) is having its own libraries and techniques to visualize large scale data but at initial stage of analysis and also for beautiful graphs we prefer to use these tools .

** Knowledge on a visualization tool is must.

9.      Handle large scale data with advance excel & spark or Hadoop:

Probably you have heard these terms such as spark and Hadoop for big data analysis. Let me clear you in this present age no companies having small scale data every odd real time data is big now a day. To process and structure these data for further analysis we must pass through an ultimate approach and techniques such as Apache spark or Hadoop.
I would say if we don’t have prior knowledge on it, needn’t to worry about that but as I suggested earlier we must open to learn it along with the practice of machine learning.

**Future is Spark, should not run from it.

10.  Knowledge on cloud platform such as AWS, Azure, IBM Watson:

Learning of machine learning concepts is different than the application of it. The real time application not only focus the accuracy but also other parameters such as scalability, reliability, maintenance and flexibility.
Keeping these concept in priorities most of the giants prefers to go for a reliable solution offered by Amazon, Microsoft and google. Hence with that note we also need to grow our skills to that level where we can use the hybrid built in models. Again, if we have no ideas on it at initial phase, need not to worry about it, learn and use of those models could be possible once you have concreate idea of basics.


**Knowledge on cloud is added advantage.



So these are the few points you should consider or evaluate yourself before make a decision to wards your career. The intention is here neither to demotivate you nor to change your decision rather then give you a clear view towards shaping your future .

These info what i mentioned are from the studies and own experience when i started my career on ML probably two years ago. it might possible in more or less i am wrong at some point, please feel free to correct me and also looking for your thoughts and doubts in comment section, i will try my best to resolve all your queries .