Saikat Basak is a Data Scientist, Artificial Intelligence Expert, Machine Learning Engineer and Author.
He is passionate about computers and modern technologies like data science and many more.
He had done his Master of Computer Applications and Master in Engineering (ME) because of his interest in computer science.
He has done a job as a data scientist in many multinational companies.
He is an author too and runs his website where he publishes articles on data science and artificial intelligence.
We got the opportunity to discuss him about data science, artificial intelligence domain and the World's Hottest job of “Data Scientist”.
He shares his journey and vision of all the latest technologies with us.
I did my graduation, B.Sc, in Physics but growing up I always felt a deep connection with computer science, almost to the level where I would stay awake for consecutive nights and do random nerdy stuff on my computer.
So, after three years of studying Thermodynamics, Relativity, Quantum Mechanics etc. I came to the conclusion that things are not working anymore between Physics and me. I had to move on. And I did move on to study computer science. I did my Master of Computer Applications and subsequently did my Master in Engineering (ME).
Around the year 2013, I started taking interest in Image Processing and Computer Vision which would later turn out to be an immense help in kick-starting my career in the Artificial Intelligence domain.
My first job was with a multinational Digital Business Transformations company (just a trendy word for consulting) which is part of a big French multinational, one of the Big 4s of the advertising industry. There I was part of an Innovation Lab and was working in the field of applied AI - mostly computer vision.
I switched to Data Science after two years and had been working in the field since, first with an online real-estate company and currently with a global online classifieds company.
Data Science sure has become the most sought after job in recent years thanks to the breakthroughs in GPU (also CPU) computing technology in the past 6-8 years. But before I start bragging about all the perks of being a Data Scientist these days, let me address the elephant in the room. And let me bust some myths while I do that.
Data Science is overhyped. Okay, that might seem a little harsh but it is to an extent highly marketed (sometimes to an extent that is misleading).
So, what is Data Science? One might say it is applied statistics, gathering relevant insights or solving any business problem using data, or just machine learning in a business suit. It is, basically, all of the above. 80-85% of a Data Scientist’s work revolves around gathering, cleaning, organizing, analyzing data. The rest is to build statistical models (or machine learning models) that give some prediction based on the data. Data Science surely sounds sexy but a lot of not so sexy work is involved in the whole process and that might demotivate some of the newcomers.
The job of a Data Scientist is to generate value for the organization. The tools we use give us the Power to generate huge value in a very short amount of time with minimal effort. It is like the Power of Grayskull if you were a Master’s of the Universe fan growing up (MOTU or any other 80’s franchise might sound alien to some of the readers, I’m talking about He-man). Data Science is the Power Sword that gives Prince Adam (our Data Scientist) the powers of the Castle Grayskull (Statistics and Machine Learning) and makes him the most powerful man in the Universe.
And it takes only very little time to generate high value for an organization if the tools of Data Science are applied properly. The tools to master are surely advanced and it takes an above average intelligence, a lot of hard work, and immense passion to become a pro Data Scientist. The scope of Data Science and the scarcity of Good talent is what makes this job the hottest job of the century.
In many ways, basically. We have to go by examples. Let me pick some specific industries, one after one, and we shall try to understand how Data Science helps those industries to solve some of their problems.
First, let me talk about eCommerce. Two words, Recommender Systems, can basically sum up a big chunk of the value making machinery is eCommerce. Recommendation is what keeps the customer scrolling and in eCommerce that have a saying, “the more they scroll the more they buy”. Inventory visibility is directly proportional to the number of sales. And there can be other examples, such as cross-selling. Finding out products that a customer can buy in conjunction with the items they have already decided to buy.
If we talk about banking or fintech (financially technology) then predicting possible loan defaulters, detecting fraudulent transactions, estimating the amount of loan one could be offered based on their past transaction activity are some of the problem areas where Data Science can show huge impact.
Instant document verification, instant verification of property damage from images are some of the things financial insurance agencies have started doing nowadays.
Pricing engines become very much essential in real-estate, or in the classifieds domain where objective measurement of price for the commodities being sold are hard.
And there are many generic examples like analyzing chat or social media posts to find out abuses, fake news or disturbing content. Building customer profiles based on past interactions etc.
It will require A Lot of passion and hard work.
One can start by enquiring what a Data Scientist actually does in their day-to-day life. And that enquiry must include their life outside of work because a lot of things go differently when you are a Data Scientist. A Data Scientist needs to constantly polish and upgrade their skill sets as there is a huge inflow of new knowledge, new tools, new competition. So before one jumps into the deep they must realize what to expect. It will be tough for you to stay relevant in this job market if you aren’t upgrading yourself constantly.
Having said that, the path to Data Science is not any secret. Rather it is one of the most democratized areas of study. Almost all big technical institutes have their AI/ML/Statistics curriculum made publicly available. And their hundreds and thousands of MOOCs (Massively Open Online Course) on the topic.
One might start with the basics of Statistics or Linear Algebra but needs to switch to a smarter learning policy after they have refreshed their 11th and 12th grade knowledge. That smarter learning policy might include choosing one particular domain of ML i.e. Classical Machine Learning (this is anyway important), Natural Language Processing, Computer Vision, ML on Structured Data etc.
Choosing one particular field will help the newcomer in staying focused in the plethora of ML content that is scattered throughout the Internet. Now, it’s time to choose some interesting problems. You want to solve the problem of detecting Cancer? Or do you want to identify Fake News? Choose a problem that interests you. Try to solve the problem and learn while you do it.
Upload your solutions on GitHub or Kaggle where the world can inspect, criticize, or use the solutions you have developed. This will boost your knowledge, confidence and resume. A quick inside note, when we look for a Data Science candidate we look for people who have some experience on working with real problems.
Your GitHub or Kaggle profile is an excellent way of showcasing your work.
Depends. As I mentioned, there is a plethora of online content out there. Choosing the one that is best suited for a beginner is somewhat challenging without proper guidance.
The rule of thumb is to go for courses put up by either esteemed universities, or renowned experts of the field. But a lot of very good content is also created by individuals who are not as popular as some others. One advice that I can provide is that, for the theory and the math go for university MOOCs or follow renowned industry experts. For the stuff that is hands-on, look for tutorials and courses by the 2nd group of not so popular individuals.
Mostly Data Science and Artificial Intelligence. I prefer writing theoretical content with an intuitive explanation of the math behind the algorithms.
For my video content though, I prefer making hands-on tutorials and courses.
Depends on which definition of AI you follow. We do not have a Artificial General Intelligence, yet. So, a Skynet like scenario would be a little far fetched.
But we are At This Moment living in an era where AI influences how we live our lives. Every purchase decision we make, the digital content we consume, things that entertain us, our socio-economic status - is influenced heavily by AI.
So, how would I see a world full of AI? I see it as the world I live in now.
Link to all my courses are available on my website which is saikatbasak.in