Raj Abhijit Dandekar

In my third year of PhD at MIT, I had a conversation with Sreedath Panat and Rajat Dandekar which changed the lives of the three of us. While walking from MIT to Harvard, we were discussing the best ways to improve the education landscape in India. Towards the end of this discussion, an idea was born. The idea was to return back to India after our PhDs and work hard for the rest of our lives to better the Indian Education system.

It seemed like a long stretch. We had not even completed our PhDs! Ideas like this come and go. But this one stayed. Two years later, all three of us rejected multi crore job offers from Amazon, Google, Tesla and Facebook. We came back to India.

Startups are powerful, but to build life changing startups with your best friends is magical. The name of our startup is Vizuara. Vizuara operates at the intersection of education, art, gamification, AR/VR and machine learning. It is the future of education and the three of us are building it right now. We aim to empower teachers to become super teachers and students to become better learners, one school at a time.

We are always looking to collaborate with passionate individuals who are keen to make an impact in the education space.

(1) If you are studying abroad, but want to contribute India’s growth story by being part of a fast growing, modern, impactful startup: do get in touch.

(2) If you are an early stage investor looking to partner with a brilliant, hard working team who is going to dedicate their lives to their startup: do get in touch.

(3) If you are an artist (designer, 3D modeler, content developer), game developer (Unity, Unreal) or a tech developer (frontend, backend, fullstack) and want to impact the lives of millions of students in a meaningful way: do get in touch.

(4) If you are a sales person who can help us build meaningful and strong customer relationships; do get in touch.

(5) If you are fresh MBA graduate looking to kickstart your career in a meaningful, impactful manner and directly work with the co-founders: do get in touch.

If you resonate with this effort, please head over to our LinkedIn page: https://lnkd.in/gxBuaBjn

The three of us are working from Pune, Bangalore and Kerala. We don't regret coming back to India for a single day!

Our website link is mentioned in the comments!

IIT-JEE Mathematics vs Indian National Mathematics Olympiad (INMO)

What is the Indian National Mathematics Olympiad (INMO)?

When I was in 9th grade, I attempted the Regional Mathematics Olympiad (RMO). It was one of the most challenging mathematics exams I had ever attempted. I could barely get a single question right.

I tried again in the 10th grade, and cleared the RMO. The next stage was the Indian National Mathematics Olympiad (INMO). It was a 3hr examination and it just had 3 questions. Even if you get 2 questions right, you move to the next level.

For 3 hrs, I attempted one problem. I couldn’t finish it. I didn’t even look at the other two problems.

30-35 students from all over India who clear the INMO attend a training camp at the Homi Bhabha Centre for Science Education, Mumbai. Out of them, the top 6 are chosen to represent India at the International Mathematics Olympiad.

How does the Mathematics Olympiad differ from IIT-JEE Mathematics?

Preparing for IIT-JEE Mathematics is like preparing to be a chess champion. You practice and practice and practice. The more questions you practice, patterns will start to emerge. If you solve 5000 questions, it is likely that the exam question will follow a similar pattern as one of them. If you recognize this pattern, you will do well in IIT-JEE Mathematics.

Does it test original, independent thinking? NO!

Why do I think so?

Just give any IITian 1 problem to solve from the Indian National Mathematics Olympiad (INMO). Even if you give them 3-4 hrs, most of them won’t be able to solve it.

This is because, INMO is an actual test of original, independent thinking. Only a student capable of original thinking will be able to even attempt half of a problem! Unlike IIT-JEE, there aren’t any patterns in previous questions you can rely on. The only thing you can rely on is the step by step thinking originating from your mind. It’s the true test of starting with a creative idea - testing - failing - trying another idea.

That's why in INMO, you have 1hr to attempt one question. In IIT-JEE Mathematics, you have 1 minute to attempt one question. In 1min, you just have time to apply past knowledge. But can you think and explore new knowledge frontiers? No.

Unlike an IITian, if a student has cleared INMO, they have learnt a very useful skill. The skill of original, independent thinking. That skill will be useful for the rest of their life.

Lessons:

IIT-JEE Mathematics teaches students the art of solving problems. Indian National Mathematics Olympiad (INMO) teaches students how to think. Once you learn how to think - you will be a much better scientist, researcher, entrepreneur, employee etc!

Rather than focusing on competitive exams, we should focus on learning skills which will be useful for the rest of our lives.

P.S: The 6 students who make it to the Indian team for the International Mathematics Olympiad (IMO) should be equally celebrated (if not more) than people who clear IIT-JEE.

I defended my PhD thesis at MIT last week! An exciting and eventful 5 year journey.

As my PhD journey began at MIT’s Julia Lab, we developed one of the very first scientific machine learning (SciML) frameworks applicable to modeling Covid-19 infections. Eventually, we showed that SciML frameworks can be used to understand effects of quarantine, early reopening, virtual virus spread and a wide range of applications.

Coming from a background of mechanical engineering, I would have never expected that I would be finishing a PhD in Machine Learning. It was a steep learning curve, but at the end, worth all the effort I put into it. I would be happy to guide, provide advice to anyone who is looking to make a transition from their field to machine learning.

Thank you Chris Rackauckas and Alan Edelman for the mentorship and support all along! Also, thanks to my thesis committee members - George Barbastathis, John R. Williams, Ali Jadbabaie and friends at MIT Department of Mechanical Engineering (MechE), MIT Civil and Environmental Engineering for their support in this journey!

A list of my publications can be found here: https://lnkd.in/g5xhtnQT

Neural networks are a bit glorified in literature.

The simplest way to think about neural networks is that they are just large functions.

Instead of sigmoid(x), a neural network does the following:

sigmoid(sigmoid(sigmoid(....(sigmoid(x)))

A neural networks just chains functions together to create a larger function.

Terminologies like neurons, hidden units and hidden layers are just fancy ways of expressing this larger function.