This month we are starting a developer series called Behind The Code, wherein we tap into the developer community in India and find out how these professionals, who are the backbone of our IT landscape began their journey and how they at the forefront of major innovations.
In our first column, we speak to Shrirang Deshpande, a Mechanical Engineer by qualification, whose career took a turn in his mid-20s after he developed an interest in automation and AI/ML. This interest led him to co-found an analytics startup, where he worked as a Business Analyst for two years before he moved to head the Data Science team for an online learning platform.
For Deshpande, even though he was moving in the right direction in his career path, the initial transition wasn’t really easy.
“That was the most difficult decision I have made up till now since understanding the data science ecosystem for a mechanical engineer was indeed challenging. When I look back, I certainly feel it was one of the best decisions I have ever made,” says Deshpande, who currently works as a senior associate for Gurugram-based Tyroo.
Presently, his responsibilities include product development, decision making and process improvement. He is also closely working with his management to device business opportunities in the space of digital marketing using artificial intelligence and machine learning
A Rough Start
With no prior knowledge in the field of Data Science, when Deshpande started his career, he had to unlearn four years of his undergraduate learning to start from the scratch of data science; which meant learning basic aspects like statistics, coding or domain expertise. Due to this, he made it a point to regularly update his knowledge on the latest libraries, algorithmic implementation and its use cases.
According to Deshpande, his four years in the industry has been that of constant learning, “I still remember when I first started studying I really liked the concepts like regression, clustering but understanding the use cases was a little difficult since you need to have some domain or functional knowledge to build these,” he says.
“Even writing codes for feature engineering and data cleaning was tough. But one thing I believed that practise and solving case studies from Kaggle will surely make this learning process easier and that is how the journey started,” Deshpande added.
What is the most important programming language?
A man who wears many hats, Deshpande is currently pursuing an executive general management program for young leaders from IIM Bangalore and he is also a visiting faculty at an analytics institute.
As an analytics instructor, it is the most common question raised by aspiring Data Scientist every now and then. Deshpande believes that Python is becoming popular among data scientists and software developers due to its ease of deploying ML models, “I think Python is the language to look for AI/ML. Even though I started my learning from R but certainly the innovations and robust libraries such as TensorFlow, Keras and even the ease of deploying ML models, Python is becoming popular among data scientists and software developers,” Deshpande explains.
Essential tools for a data scientist
For analytics: Vertical and as well as horizontal expansion of using and learning tools is important, which requires you to be perfect in some tools.
For statistics: I think Excel skill is a must have, excellence in either R or Python is important and having intermediate knowledge of SAS would definitely help.
For database: in-depth SQL knowledge is desirable along with basic knowledge of working with MongoDB. From the visualisation stand-point again good Excel skills along with expertise in Tableau or PowerBI.
Apart from these tools, knowledge of some Hadoop tools such as Spark or Hive is certainly good to have.
What’s in his developer kit
For statistics: I have expertise in Excel & R trying to get a strong hold on Python and important python libraries,
For visualisation: I use Tableau and for database management SQL.
Apart from this I keep on learning and reading about other tools such as Alteryx, PowerBI, HIVE, Spark Scala etc.
Word of advice
Practise, focus and self-belief.
“In India, this field is opening up quickly, I think in the next three years we will experience exponential growth in hiring and upcoming data science professionals who have good coding skills, stats knowledge and domain expertise have a really good chance of riding this wave,” Deshpande says in an optimistic tone.
What’s In Store For The Future
Learning is a life-long and Despande exemplifies just that. He intends to hone his analytics as well as his business skills in the coming few years. “I think I have a lot to learn, right from image recognition to advance speech recognition since I feel these technological advancements will be major forces driving innovation in next 5 years,” Deshpande explains