STATEMENT OF PURPOSE

We have always been mimicking nature. The Wright brothers were inspired by flight patterns of
birds to build aeroplane, RADAR detection was inspired from the way sharks identify their prey,
sonar from bats, bullet train from beak of the kingfisher bird and recently neural networks which
mimic the neuron in our brain was introduced. These show human beings are really good at
scientific observation and are bound by it. Analytics has been an integral part of the stock
market and businesses for some time now. We have been identifying trends in data like moving
averages, exponential smoothing, Holt’s Winter model since the 1960s. But we are very limited
by this kind of linear thinking. Data Science provides us the means to automate any task, given
enough data to learn. This is the next step of our evolution, turning our confined thinking to the
limitless capability of an AI which is at our fingertips to grab. Now, we have the infrastructure
and resources which make this kind of linear thinking obsolete. The introduction of
sophisticated machine learning algorithms such as Support Vector Machines where we convert
a non linear problem to a linear problem in higher dimensional plane is a great attempt which
shows when we are out of depths, we think of a way to bring that back to a familiar domain to
analyze. Random Forest, Gradient Boosting, Neural networks have completely transformed
Analytics as we know it. We are at the verge of an AI revolution which entails intelligence
explosion with booming areas of study such as Natural Language Programming (NLP),
Computer Vision, Predictive analytics.
My physics lecturer in high school, Mahender Pal Singh is good at making us visualize
the concepts we learned with real-world examples. He was very passionate about physics and
always invited any crazy questions we had. He simplified rotatory motion and Electricity topics
which we were struggling with at the time. He encouraged me to pursue Electricity as my major
in University. I love to be more practical and wanted to test out theories I have been learning on
micro-controllers, electric motors, electromagnets, etc. I worked hard to clear the Indian Institute
of Technology Joint Entrance Examination(IIT-JEE) and joined IIT Delhi, Electrical Department. I
expected the university to take a more practical approach to teaching. In the first two years, we
spent so much time on theoretical mastery that we didn’t have hands-on experience until my
third year. I worked on many good projects like Wifi clock- Synchronisation of clocks across the
university using ntpdate and iitd internal time server, Open Source Engine Control Unit(ECU) for
Internal combustion Engine, Digital Card dealer and Wildlife Camera trap. My team was
awarded the first prize for the best project in Electrical Department for the year 16’-’17 for the
project on Wildlife Camera Trap.
Courses like Introduction to Computer Science, Data Structures and Algorithms,
Computer Architecture, Computer Networking during undergraduate degree and practical
experience with embedded software based projects honed my skills as a Programmer. I have
completed a course by Andrew Ng on Introduction to Machine learning which changed the way I
thought about my career. I regret the time I wasted not coming across such an awesome field
sooner. So, I immediately started my career in Data Science after graduation. I started data
science course tracks on coursera like Deep Learning by Andrew Ng, Data Science Toolbox by
Johns Hopkins University, Big Data by University of San Diego and Applied Data Science by
University of Michigan. I finished introductory courses in each track and am currently learning
the second round of courses in each track. Geoffrey Hinton’s interview on how he started his
career in neural networks when there was no foreseeable hope for the field and the way his
team made it to the leaderboard on Imagenet data on Kaggle using convolutional Neural
networks(CNN) inspired me. I followed the interviews of various experts which pushed me to
further explore data science.
I’ve been working as a Software Engineer at GGK Technologies for the last 10 months
with a focus on Data Science. I worked on predictive analytics for a project called Luminat, a
Real Estate Advisory Company. We, as a team, designed a machine learning pipeline for
preprocessing, feature engineering and modelling. Features we used included the data
collected by realtors and enhanced with the data purchased from a third party client called
Experian which provides Credit based statistics. I’ve also worked on time series modelling for
revenue forecast for a client to estimate the monthly revenue accurately for the valuation of the
company accurately. ARIMA came out to be the best estimator for revenue.In addition to these, I
worked on text analytics for a supply chain management company to keep track of spelling
mistakes and duplicate company names. I’ve learned a lot from my work, which I think will help
me both professionally and personally. I wanted to explore more about word embeddings for
name matching, deep learning for the Luminate project, video analytics for warby parker but the
company’s interest goes against my research interest. My dream is to build a General purpose
AI and I believe Masters in Data Science is a step in right direction.
My Research Interests include Computer Vision, Digital Image Processing,
Convolutional Neural Networks(CNN’s), Pattern Recognition, Recurring neural
networks(RNN’s), and GAN’s. I am still participating in Kaggle competitions, pursuing Coursera
courses and going through the data sets and the approaches people are talking about. I would
like to contribute more to these areas of research.
I see tests conducted by Andrew Ng’s Drive AI on self-driving cars in Texas and it is
inspiring to say the least. The US undoubtedly leads the rest of the world in terms of AI
research. I would like to be at the technological spearhead of our era to learn from the best.

\\Northeastern
The reason for choosing Northeastern University is because of its unique Educational
Strategy. The Co-op, experiential learning part of the program appealed to me. It would be great
to develop some real experience in data science. I would love to learn more about research
conducted by Associate Professor Michelle Borkin on Human-Computer Interactivity and
Visualisation techniques she has been working on.

\\USC
The reason for choosing the University of Southern California is because of its outstanding
research in AI, especially in Computer Vision. I am interested in projects such as 3D modelling
of faces through video, use of knowldge in vision(veil) and automated video surveliance using
moving object detection. I will receive the education and technical skills I need, given the terrific
academic reputation, rigorous challenging courses of the University of Southern California.

\\University of Maryland
The reason for choosing UMASS is because it has a great curriculum with an emphasis on
practicals. I would love to learn more about research conducted by Professor Dr. Ramprasad
Balasubramanian on visualization techniques, pattern recognition and Human-Computer
Interactivity he has been working on.

\\UIPUI

The reason for choosing IUPUI is because it has a great curriculum that is a mix of statistics and
machine learning which will prepare us for any challenges in a work environment. In research
terms, I read a paper on Identifying Natural Events through Social Media and Computer Vision
which is a novel idea to use daily streaming data for noble purposes. The University of Michigan
has done something similar using the Twitter feed to identify fire disasters. I can pursue
computer vision research and hone my technical skills given the terrific academic reputation,
rigorous challenging courses of IUPUI.