Machine learning study sop
It seems almost magical. Saying something to a screen and hearing something in response. Getting your ETA for office even before leaving your house. Under all these technologies, machine learning is at play. I believe that machine learning plays a very big role in helping humans expand their capabilities and create new possibilities. From giving you recommendations on what to buy on Amazon to the digitization of medical samples in rural African villages for the efficient provision of healthcare through character recognition, machine learning is broadly integrated with the ways we interact with technology and our society today.
Machine learning owes its evolution to a massive amount of data generated in recent times and advancements in computer hardware. Understanding this massive amount of data then becomes a necessity. I have been involved in data exploration since my junior year. I started with exploring GitHub data which included commits of repositories that used Travis CI. I aimed to explore the correlation between the number of team members versus build failure in a GitHub project. Ultimately with a few refinements to my procedures and results, I published these findings in a journal. I got introduced to Machine Learning when I was tackling the issue of intelligently parsing the textual data during my internship. I quickly realized how mathematically and programmatically involved the field of machine learning was. Since my freshman year, I have thoroughly enjoyed programming and decided to take as many Algorithm and Programming classes as possible. By the junior year, I had a strong foundation in algorithms, data structures, and programming. I delved towards the mathematics of Machine learning and with the help of various online courses and books, I developed an understanding of building blocks in Machine Learning algorithms. At the same time, I maintained a great academic record in my electronics and communication engineering and finished among the top 5% at my university branch. During my master's degree at the University of Maryland, College Park, I hope to contribute to the research relating to intelligently mining, learning from data and building robust machine learning solutions.
Towards my senior year, while doing my introduction to image processing class, I got interested in computer vision and how deep learning algorithms have advanced to solve nuanced problems in the field of image processing. Thus, I took 'Detection for Traffic Rule Violation using Object Detection' as my undergraduate project where I discussed various algorithms used for object detection and how I have implemented to solve the challenge of traffic rule violation using YOLO algorithm. This turned out to be a very substantial learning curve for me where I studied various algorithms in the domain of computer vision. While working on this project, I focused on learning the fundamental concepts and work my way up in building a neatly engineered solution. It laid down the foundations in me for a research-oriented approach towards machine learning and I consider my undergraduate to be an introduction to this field.
During my final semester, I got selected as a Data Science intern at <SaSS company>. I soon got converted as a full-time employee and has been working since then at <SaSS company>. At <SaSS company>, my learning has been steep and very rewarding. I have been involved in the whole lifecycle of data and have learned a lot about databases, social networks, and modeling data at scale. One of my favourite project at <SaSS company> was modeling the user network on the graphical database and finding inferences from the bipartite graphs formed which ultimately led to a product decision of making a core feature available for free on the app. I have also been responsible for creating Data Science APIs and propensity modeling to study user usage patterns. The work culture at <SaSS company> has enforced my thought process for data-driven decisions, and architecting engineering solutions at scale. I also improved my analytical thinking, my capability to work independently as well as in teams, and under pressure, as I directly report to CTO of the company.
Many research projects at UMD are interdisciplinary in nature. I believe that my standpoint has also developed in a similar fashion - sound knowledge of communication systems, signal processing, programming, and machine learning. This helped me in qualifying for the semifinals of ARM Advanced Computing and Communications Society - student design challenge 2016, where I was required to submit a solution of smart traffic control using IoT. I was also appointed as a TA for Data Structures and Computer Programming Lab. I hope to continue this at the UMD too. I volunteered as a proposal reviewer for Google AI Impact challenge 2019 and reviewed 14 promising applications based on the social impact from a pool of 2000 applications. While reviewing those applications, it was clear to me that Machine Learning and AI can drive a lot of good changes in society through technology.
I have always considered that in the industry, while you are applying any technology, you are using them as a means to an end to create business value. However, research pushes the boundary of these technologies itself to new limits. I consider research more liberating as it needs a theoretical and analytical approach. This is the core reason to pursue a master's degree. My primary field of research interest is machine learning models in the domain of computer vision and probabilistic graphical models I am excited about the ongoing research at the UMD, particularly in the field of graphical learning models by Dr. Danai Koutra at Gems Lab and the research being done by Dr. Honglak Lee in the application of machine learning to computer vision. Given my industrial knowledge and undergraduate university education with the highest grades in the related coursework, I am in a good position to contribute towards the research under the guidance of the esteemed professors. I firmly believe that going through an intense master's degree program with the concentration track for machine learning at the UMD would give me immense exposure, much deeper knowledge of the domain, and aid me towards better understanding the theory and application of machine learning models at the intersection of research and industry where systematic data-driven decisions can be made.
