In this changing ecosystem, we will see mass democratization of technologies like AI/ML. These technologies will solve some of the biggest societal challenges in the coming years, including the eradication of a lot of diseases we know today, cleaning up the oceans, exploring new planets to protect people during disaster situations, etc. Kunal Jain, CEO and founder, Analytics Vidhya, tells us more.
DQ: Give us an overview on Analytics Vidhya?
Kunal Jain: Analytics Vidhya aims to create next-gen data science ecosystem across the globe. We are India’s largest community of data science professionals and aspirants. With the rise in amount of data generated across the globe and businesses wanting to harness latest technologies like artificial intelligence and machine learning for their businesses. The demand for data science professionals is continuously increasing.
When I started Analytics Vidhya in 2014 – there was no ecosystem in place for professionals to interact and learn from other people in the domain. We used to rely on our own natural circles to find answers to these questions. Over the last 6 years, we have helped millions of people learn data science and unlock these unique opportunities in their career. Thousands of people today use Analytics Vidhya for their daily learning, taking courses, competing for some of the best opportunities in domain, and getting jobs based on their skills.
DQ: How can analytics / data science bring about a significant change in the current scenario?
Kunal Jain: We are currently living in a world that is already being driven by the use of data science and ML. The voice assistants we use today for the cameras in smartphones are driven by these technologies. The latest spacecrafts use ML to identify their landing spots, to charting their own territories.
We will see mass democratization of these technologies in the next 10 years. Imagine children learning about these tools and techniques in schools and colleges like we studied coding. They will be enabled to create their own smart universe using autonomous vehicles, high-quality personalized healthcare to smart drones and nano-bots, and automating things which need human interventions today.
These technologies will solve some of the biggest societal challenges in coming years, including eradication of a lot of diseases we know today, cleaning up the oceans, exploring new planets to protect people during disaster situations.
DQ: What are the most important attributes to have for a data scientist? What is the potential for the next generation workforce in this field?
Kunal Jain: At the end of the day, a data scientist is a problem solver using data to solve problems in business and society. In order to handle the data at huge scale, you obviously need the technical skills like mathematics, statistics, data engineering and coding. You also need the right soft skills, like communications and co-ordination with your stakeholders and the domain knowledge. Until you understand why things are happening in a particular manner and how does that translate into data, you will not be able to make sense out of it.
To summarize, the three broad set of skills required to be a good data scientist are excellent technical skills, good soft skills and deep domain expertise.
These tools and techniques would open up new avenues in several domains, some of which we are not even aware of. For example, I see an expert responsible for AI ethics in every large organization. This role does not exist today, but would become critical for the continued success of AI/ML initiatives in several companies.
DQ: What are companies looking for when hiring data scientists and ML engineers?
Kunal Jain: We have seen a lot of innovation on this front in our community. Since, there are only a limited number of people with the skill sets, companies have tried several non-conventional ways to hire talent from the community.
For example, we have seen some of the top organizations come out in open and conduct competitions and hackathons for hiring. These hackathons provide a unique way to hire as companies can see candidates solve problems before hiring them, and community members, who may not have formal education in this space, can prove themselves and get hired. Going forward, we expect more and more companies to take up these unconventional methods of hiring.
DQ: How do you think about data before implementing tools? How can one know if a data science project is worth continuing?
Kunal Jain: Now, this might sound a little ironic coming from a data scientist, the truth is that we do not think about data before solving problems. The problems that a data scientist works on, should originate out of strong business needs. Once these business problems are identified, they can be converted into data science problems.
For example, if an e-commerce company wants to increase the lifetime value of a customer (business problem), they can do this by creating a recommendation engine, which tells the likely products a customer will buy based on their profile, past purchases and browsing. Once this is done, that is when you actually start thinking about data, not before that.
Further, almost every big data science project happens iteratively. You build simple models to start with and then improve your models as you gather more data. For example, the first version of Google search was the same for everyone. Today, Google search provides individual recommendations and search results based on our history. As long as you think there is a big business value to be unlocked from a data science project, you continue to work on it.
DQ: What are some common data science questions? What does interview process look like?
Kunal Jain: This depends a lot on the nature of the role. A typical interview process would include multiple rounds assessing the candidates on the technical skills, fit with the team, case studies and role plays. Some companies also use puzzles and guesstimates in their interview process.
I personally think that any format, which replicates the real life problems, which the person would be working on, is the most effective format. This includes case study discussions, model building assignments and hackathons, to name a few.
DQ: From a talent point of view, how can data science and analytics bring more women into workforce?
Kunal Jain: Very good question. It is something I personally spend a lot of time thinking about as well. We need to have a diverse representation of talent in order to set it up for long-term success. To that effect, we have been encouraging more and more community events among women. For example, we recently conducted a women-only hackathon that led to multiple hiring with one of our clients.
We actively encourage more and more women to come forward in community activities and act as role models. We have constantly showcased women role models and their work to the community in the form of podcasts and several other initiatives. I think, we are on the right track. The problem can only be solved using wider representation and engagement.