DSAT – First Ever Adaptive Learning Platform for Data Science Professionals

Tavish Srivastava 03 Dec, 2019
5 min read

“Every once in a while, a revolutionary product comes along that changes everything.” – Steve Jobs

We are thrilled to announce the launch of the first-ever adaptive learning platform for data science professionals around the world – DSAT, or the Data Science Adaptive Test!

Here’s one of the most common scenarios we have seen in the data science industry:

Hiring Manager – “It’s been 6 months since we opened up this job position. All you have got me are at best round 1 rejects. We have already lost more than $1 Million in opportunity.”

Human Resource Partner – “Your demand is just too specific. You need a data science professional with an understanding of the banking domain, a master of the digital marketing function and ability to work with both Big Data and unstructured data. I have already given you my best leads from the pool. All the future leads will be even worse. You need to relax your criterion to close this position.”

Believe me, this is not an exaggeration. I have seen far more aggressive discussions and frustration in the hiring process in the data science industry.

With domains like Data Science, Cognitive Tech, Information Tech, Software Engineering, Semiconductors, Data Architecture, etc. all converging to solve real-world problems, jobs are increasingly becoming more skill-demanding and recruitment is becoming quite a complex problem to solve.

This learning and recruitment challenge, combined with about a 20% quarterly increment in job openings, accounts for huge economic losses across industries.

So how can we cope up with such an increasingly complex challenge of learning and recruitment? To answer this question, we need to take a step back and zoom out.

Note: If you prefer visual to verbal, here’s a quick video explaining DSAT, the logic behind it, and why it’s a revolutionary product for both data science professionals as well as hiring managers:


Reflection on Historically Solved Problems

Back in 1953, every business school was facing the same challenge of low application pool, a poor and extremely operations intensive selection process, and a very difficult time establishing a comparison of selected candidates across these schools.

Necessity is the mother of invention – the top 9 business schools including Harvard, University of Chicago/Michigan/Pennsylvania, etc. met for this burning need and established a state-of-the-art admission test that could solve all these challenges at once.

But how can a single test solve everything?

It all boils down to quantitative metrics that can be easily measured. What used to be a top 4 city selection process can now scale across the United States, penetrating even rural areas. All the candidates are scored on a single metric that can judge how good/bad the candidate is.

With this score, the fierce battle to get the best candidate came in the competitive market instead of the agency model that existed. In 1997, they even solved the only challenge left in this admission process – evaluation across the entire pool through a computer adaptive test – a state of the art technology at that time.

Unlike fixed tests, this computer adaptive test can distinguish every candidate from each other regardless of their position in the crowd. Fast forward 67 years – now 250,000+ tests are taken across 100+ countries. What made this test so popular? Standardized Adaptive Test on globally agreed subjects!

The same challenge was solved through GRE in other educational domains, through CAT in the Indian Business Education industry, through SAT in US undergrad admission, etc. None of these exams is pure computer-adaptive yet – but all of them strive to get there someday.


What Makes Computer-Adaptive Tests so Difficult to Build?

There is a significantly complex process that goes into creating and designing a computer adaptive test.

A simple (fixed) test of 40 questions will need you to answer those 40 questions only. But an adaptive test of 40 questions will need at least 400 questions with various difficulty levels. Tagging these difficulty levels will in-turn need at least 1000 attempts per question. So, you basically need 400 questions with about 4,00,000 attempts from more than 1000 individuals.

This should give you a good flavor of what it takes to make an adaptive test and why so many standardized tests have not scaled up to being truly adaptive.


So the question we wanted to answer was – can we bring this intelligence/experience in our industry to solve our challenge of learning and recruiting?

We rolled out a product a few months back called Datamin to validate if a single product can solve both the learning and recruiting challenges of the industry. Long story short – the answer is YES!

We believe that a single product similar to GMAT can revolutionize this entire industry.

After the successful launch of Datamin, we wore our community hat and thoughtfully built a product for our community that could do both. Here is what our community was able to accomplish with Datamin:

  • 10,000+ Users
  • 500,000+ Questions Attempted
  • 99% Questions now tagged as unambiguous

Imagine the kind of intelligence we were able to create. We used this intelligence to upgrade our plain vanilla but addictive Datamin to the first-ever data science adaptive test – DSAT. DSAT stands for Data Science Adaptive Test.


How will DSAT solve the existing recruitment and learning challenges of the Data Science industry?

We built an illustrated infographic to explain:



What skills does the DSAT test?

There are primarily four skills you will work on (and be evaluated on) in the DSAT. Here’s the lowdown on each skill:



How will the DSAT evolve over time and how useful will it be for you?

There are multiple avenues the DSAT will open up for you. Here are two broad ones that we want to highlight. At the very least, you get to work on your data science skills, evaluate them, understand which areas need improvement, and then work on them:


The other scenario, catering to everyone from aspiring data scientists to hiring managers, is all about getting noticed by potential data science recruiters. Work on your DSAT score, highlight it to hiring managers, and stand a chance to land your dream data science role:


End Notes

DSAT is a product which is truly democratic – for the people, by the people, and of the people.

We have done our job by creating the underlying technology that can foster a higher degree of learning. Now, it’s your turn to use it to the fullest.

The score is as good as its adoption. Any community is as good as the people within. Hence, make sure your DSAT score reflects your true potential, and if you believe this score does justice to your knowledge, make sure to bring in your peers to have more credibility to your score.

Get started with DSAT today!

Tavish Srivastava 03 Dec, 2019

Tavish Srivastava, co-founder and Chief Strategy Officer of Analytics Vidhya, is an IIT Madras graduate and a passionate data-science professional with 8+ years of diverse experience in markets including the US, India and Singapore, domains including Digital Acquisitions, Customer Servicing and Customer Management, and industry including Retail Banking, Credit Cards and Insurance. He is fascinated by the idea of artificial intelligence inspired by human intelligence and enjoys every discussion, theory or even movie related to this idea.

Frequently Asked Questions

Lorem ipsum dolor sit amet, consectetur adipiscing elit,

Responses From Readers