A curious child, electronic engineering occupied the center of my world very early on, taking old electronics devices apart and putting them back together again, exploring. Electronics was one of my great passions that I pursued to early adulthood, earning my BS Degree in this area. Despite the fact that I majored in Electronic Engineering as an undergraduate student, I have also excelled at Math and Physics and long cultivated a special passion for these areas of study as well. Gradually and for some time now, despite earning my degree in Engineering, I have come to believe that my professional calling lies in the area of Statistics.
The bridge for me between electronic engineering and statistics was my close relationship with my computers, gradually assuming a central role in my life becoming an integral component of my professional mindset and spheres of interest. A good student highly motivated to the achievement of excellence in education, the MS in Statistics Program at Stanford University is my first choice for graduate school for several reasons, most of all the sheer excellence of your program along with its location. I keenly look forward to giving my all to distinguishing myself, in particular, in the areas of Statistical Learning and Optimization. I very much admire the thoroughly interdisciplinary nature of the program at Stanford with the flexibility of choosing electives from a wide range of domains, including engineering, and I am also keenly excited about the prospect of learning from and contributing to the consulting services offered by the Statistics Department. Statistics for Good is also an initiative by Stanford that I very much admire since volunteer work has been close to my heart ever since my days as a volunteer teacher in Varanasi, here in India where I have been born and raised. I deeply respect and would love to be a part of Statistics for Good in any way possible. If accepted to Statistics at Stanford, I will most definitely attend the program.
I see myself in the future as a member of an industrial research and development team working on developing statistical learning and optimization algorithms to solve problems in engineering.
As I see it, earning the MS degree in Statistics is the logical culmination of my passion for statistics and optimization in general. I hope to be accepted to Stanford because of its world-class reputation and research atmosphere, the best faculty, exceptional peers, and excellent infrastructure, the optimal location for exploring my potential. A high achiever right since my high school days, in 2012, I was awarded the Best Student in the City of Mumbai Award for high academic achievement and distinguished extracurricular activities. Similarly, I was awarded the most coveted Director's Gold Medal for outstanding all-round achievement and leadership among some nine hundred B.Tech graduates of the 2018 batch.
Along with my Engineering courses, as an undergraduate student I also studied Math, Probability, and Statistics, with the latter stealing my heart. In my sophomore year, I served as an undergraduate researcher in the Sensors and Systems Lab at IIT BHU, where I worked as an undergraduate researcher on the classification of gases using the response data generated by the gas sensor array. This represented my first major exposure to data preprocessing, machine learning and swarm intelligence. I took the much-vaunted online course: Machine Learning by Dr. Andrew Ng, to get familiar with the basic concepts quickly. In the end, I developed a CNN model from scratch and trained it with PSO to get a faster convergence; reducing the dimension of the data using PCA to tackle the highly correlated sensor responses. Learning a great deal about statistical learning and optimization, I was also given a glimpse of how core engineering problems might be solved as a result of the discovery of hidden patterns in data.
I now have a solid academic foundation for further study in optimization and statistics, especially digital signal processing, forecasting, and Time Series Analysis. My courses in quantitative methods as well as information and distribution theory and wireless communications have been balanced with courses involving a lot of programming, mathematical modelling and simulation. I also have a solid grasp of artificial intelligence and parallel algorithms. Linear programming, solving PDEs, Bayesian learning, Riemann Integrals, queuing theory, MDPs, game theory, neural networks, fuzzy logic, pdf, to name several, are all integral parts of my professional toolkit.
Having Dr. XXXXs, an Assistant Professor at IIT, BHU, as my advisor was a special honor, as he was perfect for helping me to develop my focus on learning and optimization along with the Beam Selection problem in massive MIMO systems for 5G communications. We explored traditional techniques like Markov Decision Processes, maximum weight matching and found Matching Theory, a concept predominantly used in Economics, uniquely suited for the research. We modeled the problem into a matching with externalities game and devised a novel heuristic algorithm to get a stable result. Then, we also came up with a novel linear precoder based on QR decomposition of the channel matrix, further improving the results, significantly better than the state-of-the-art algorithms, resulting in publication in IEEE Access.
It was also a special honor to serve as a Teaching Assistant for the course 'Probabilistic Graphical Models' which involved tutoring students after class, answering their questions and helping them to formulate solutions to assignments as well as grading papers. Dr. XXXX, currently working at Ericsson Canada, has played a major role in my decision to pursue a career as a statistician. I still work with him on finding DNN based approaches to solve problems in 5G communications. I have been affiliated with NVIDIA both as a summer intern in my senior year and as a full-time engineer after graduation. I form part of a team which develops a tool that projects power and performance numbers of upcoming chips using statistical techniques. In the past year, I have also worked on product binning analysis, a convex optimization program that allows large variances in performance of the finished chips to be optimally condensed into a smaller number of marketed designations. I am also currently working on developing non-linear regression models to project the leakage power and clock numbers of next generation GPUs. The remarkable thing is that the tool assists higher management to make crucial business decisions and I have seen instances at NVIDIA where decisions with millions of dollars hanging in the balance have been largely made based on the data generated by this tool. For me, this underscores the incredible potential impact of statistics.
I thank you for considering my application to Statistics at Stanford.