Discovering the Science Behind the Data Scientist : In Conversation With Jacob Joseph
Company

Discovering the Science Behind the Data Scientist : In Conversation With Jacob Joseph

Have you had the opportunity to work alongside award winners? We’re privileged to work with Jacob Joseph, our lead Data Scientist, who was recently included in the 40 under 40 Data Scientists in India by the Analytics India Magazine.

We chatted over coffee about the award, his background, some cool projects & words of wisdom. Read on :

Firstly, hearty congratulations Jacob, on making it to the prestigious 40 under 40 list. Tell us more about the award.

Thank You Pooja. When it comes to age being a factor of assessment, I must say I made it in time to the 40 under 40 list! On a serious note however, it sure is a great honor to have been nominated & selected as a winner among some of the smartest brains in the Indian data science community.

The award itself is prestigious as the criteria was not only about having personal achievements, but also a prerequisite to have delivered substantial business impact & contributed to the advancement of the data scientist role and to the analytics ecosystem in India.

The jury comprised the editorial team of Analytics India magazine, industry experts, academicians & business heads of reputed establishments; and it feels absolutely great to have been awarded by this community.

That’s great to know! We’re all extremely proud of your achievements! But let’s start from ground zero – tell me how it all began.

Coming from a commerce background, data science was a far fetched thought in the beginning. After spending a decade in investment banking, venture capital and consulting on mergers & acquisitions, I realised that fund management was what interested me the most. Using mathematical models to price assets in a quant-based investment style became my passion.

I quit my job and focused my efforts on improving at math, statistics, and programming. I completed courses on financial engineering, quants, & analytics. I kept utilizing my learnings on freelancing projects & participating in hackathons, one of which I won. In 2015, I joined CleverTap as Data Scientist and was one of the first 2-member data science team here.

That’s one interesting journey! I’m curious: what distinguishes the top 40 data scientists from the rest? Who makes the cut?

I believe that all the other data scientists who made it to the list must have also tried to understand their consumer problems and developed solutions with the help of data science.

What has always worked for me is the problem solving approach towards things. Any data science project begins with a problem statement.

“In addition to speaking to the team at CleverTap, I make it a point to speak to at least one customer a week to understand their problems and suggest solutions that CleverTap already possesses or could build.”

 

It’s a customer’s problem statement that motivates me and my team. Here we believe in the co-creation framework and every data-science feature is launched with an objective to solve a real customer problem.

This is really interesting! Can you quickly tell me about one of these cool problem-solving data science backed features that your team worked on and how it came about.

When I speak with marketers, some of the most common issues they face are: how to drive KPIs; or how to arrive at better data-driven decisions faster, or understand the root cause of a problem.

Even though the problem statement itself is not very well defined, this is a trigger for us, to find a way to help marketers overcome these issues.

By delving deeper into their conversations & putting on the data scientist hat, we understood that marketers are mostly looking to reduce uninstalls, increase revenue, decrease churn, and more.

So what they really wanted was someone to give them the right consumer segments to target (campaigns & communication) so that marketing budgets aren’t wasted on people that would either buy from you anyway or worse, never buy from you!

To do this, CleverTap’s Coeus Data Science Engine launched Intent Based Segmentation and combined it with predictive modeling to precisely identify the users most likely to perform an event in the future.

Marketers are definitely thankful to data experts like you! But who are the experts you look up to?

There are some great masterminds who have served as role models for me. Among them, Geoffrey Hinton, Andrew Ng, and Yann Lecun are my exemplars because not only have these individuals excelled in academia but successfully transitioned to make an immense impact in the business world.

What would you advise marketers when it comes to dealing with data?

In today’s ultra-competitive world, growth marketers spend valuable time and resources to get going with their campaigns, messaging, targeting and more. My word of advice would be to understand data without having a bias.

It is great to be open to automation, machine learning, and use tools & models to serve you results, however, one must never embrace it blindly without understanding how these models work in general.

 

Even our Marketing Head, has always believed that ; “Without an opinion, you’re just another person with data.”

What will data science look like 10 years from now? What would your advice be to budding data scientists?

“Data Scientists are working really hard to put Data Scientists out of their jobs!”

A lot of the stuff we do – like data cleaning, exploration, and even selecting the right model for the problem – is getting automated. And this trend is only going to increase. With the advancement in quantum computing, the model building time is also expected to decrease drastically.

My guess is that data scientists 10 years down the line will focus more on ensuring the quality of training data that goes into the model, possibly the data science architecture, and understanding and getting deeper insights from the current black box models.

What I would tell anyone thinking of getting into the world of data science would be:

Cultivate the art of being patient as results can be delayed. Do not jump into model building when you see data. Understand the problem and explore the data very well. There are no shortcuts.

“Never fall in love with your machine learning model.”

 

Adapt as per the problem you are trying to solve.

Also, for those who plan to take this up as a career, it will be good to have an industry agnostic approach. Work with companies like CleverTap who serve customers across various industries such as fintech, media, food, travel, ecommerce, and more! You’ll learn something new every day.

8,000+

$2 Billion

1 Billion

20 Million