Profile

I am an amateur photographer and doodler. Sometimes I do a bit more serious painting to express myself. Whovian. South Park fan. I love cooking and swimming. I am also known for getting hurt too often due to my hobby of jiujitsu. Once upon a time, many many years ago I had trained to become a singer in Indian classical music. Then things arrived at a point when I had to choose between music and science. I had to quit music. Hence now that I am independent, after 12 years of separation from practicing music, I am trying to find my way back to it, but this time solely for myself. I sing while cooking and recently I have bought an ukulele and I am learning (more of struggling) to play that. I am told that I have a very dark sense of humour. I love solo travels. Whenever work and surroundings become too confusing and suffocating I plan and go on solo travels.

I was about to throw away the IKEA catalogue but then this happened

There was this depressing phase earlier this year (due to research). I was crying uncontrollably and couldn’t find the words to express what I felt. That’s how this digital painting was born. It is inspired by the song Asleep, by The Smiths.

My team and I have developed a ‘machine learning’ program which can identify rare diseases of the adrenal gland (a cap-like gland on top of each kidney), and help doctors find out faster which part of the adrenal gland is not working. The faster the doctors are able to find the problem the faster they are able to treat it.

Machine learning is where we teach a computer to recognise patterns in big data sets. In this case, the computer learns about perfect examples (‘prototypes’) of different adrenal diseases, from data about patients’ heart rate, blood pressure, ECG profile, and urine and blood test results. We can then show the computer some of these data from a new patient, and the computer can tell us which disease this patient might have.

Sometimes while dealing with rare diseases, doctors often do not know which hormones in the patient’s body are causing their symptoms. The program my team and I have developed can also give more information about which hormones are more and which hormones are less responsible for the disease. This can help the doctors to understand the adrenal disorders better.

We use computer science, maths and physics to help build up the model and make sure that it is as accurate as possible.

And here is a gif which shows how I use Geometry to make results of grouping by my model to appear on a spinning sphere (which I have named Sphere of Doom). The stars show the prototypes which are, as I mentioned before, the ‘perfect examples’. The alpha-numeric codes  are the names of the hormone disorders which my program can find and gain knowledge about.

I have recently bought a ukulele. Everyday after waking up around sunrise, brushing and getting the first cup of tea, I try to learn how to play the ukulele less horribly than the previous day. After an hour of practice I get ready to go to my office where I spend the day trying to find out patterns (prototypes) from numbers generated by big medical machines. I teach the computer to learn to spot these patterns when a new measurement comes from the big medical machines.

My office corner 🙂

Late evenings some days I go to my dojo for jiujitsu (currently training for green).

Other days I just return home to cook food (I cook just twice a week, I am a very lazy person). Sometimes there are fun things to do with friends 😉

Karaoke night at a friend’s place

If I win the money I would like to organize science communication trainings for fellow PhD students and post-docs. I believe that unless we as scientists are able to explain our work to different people, of all ages, and educational background, our work in science is only half done. With the encouragement and feedback I get from you I think I will be ready to explain to more doctors what machine learning can do for them… So I’d organize a meet up session with the overworked but sweet doctors in my city in India and tell them about what machine learning can do to ease their work load. And since it is because of school going brilliant minds that I would have been encouraged this much I’d like to visit a few schools in my city to talk to the students there and tell them how Maths and Physics are so much more than just getting good grades and securing a place in the prestigious universities.

Do science blogging in such a way that everyone can understand what I am trying to do.
My fellow scientist friends and I had once planned to make songs on what we do, such that students find it fun to learn and they remember better. Probably these last 2 don’t need much money, but just sharing an idea of what else I intend to do in the near future for science communication.