Sample SOP – Computer Science
First Draft
To explore the role of AI in recognizing human emotions, apart from its cognitive understanding ability, in order to analyze human sentiments better, I worked on a project called AI Therapist. This project was aimed at developing a system that can recommend therapy based on the user’s emotions/mood, which the system would detect using either video/image, audio or text. For emotion recognition using image or video, I used Computer Vision, Face Detection and trained a Convolutional Neural Network (CNN) for emotion recognition using video frames. For emotion recognition using audio our first approach was using speech to text and a BERT model, but the model couldn’t understand the intonation of the speech, which meant that speech to text would not work in all circumstances. After some research, I came across Mel Frequency Cepstral Coefficients (MFCCs) which is very close to how the human auditory system senses audio and trained a CNN using the same. Additionally, I also provided a chatbot feature for the user to talk to where I used a neural network for chatbot and BERT model for emotion recognition using text.
The above project sparked my interest in NLP, and led me to work on Exam Preparation Kit Using Natural Language Processing where the objective was to help the students with their preparation for the exam by allowing them to say their answers into the system which would then be evaluated. The system asks questions based on the previous performance of the students i.e. asks from their weak topics. After answering a question, a score along with a reasoning breakdown of the score and explanation is displayed. The main challenge was evaluating the student’s answer using the right answer given by the professor for which I used a Siamese BERT model that perfectly fit our use case and gave us better results. Through this project I learned some important concepts of NLP such as tokenization, vectorization, summarization etc. I authored a paper on the implementation of the project which is awaiting publication in IEEE and presented at the 12th ICCCNT held by IIT Kharagpur.
My research inclinations made me work on additional research papers. The paper titled “Airplane Crash Severity Prediction using Machine Learning” (awaiting publication in IEEE and presented at the 12th ICCCNT held by IIT Kharagpur.), used several machine learning algorithms, a neural network and stacking ensemble technique to predict severity of airplane crash based on factors like weather, engine type of aircraft, phase of flight, purpose of the flight, etc. In this paper, I also compared the algorithms used and justified the performances. Max accuracy achieved was 91.66% by the stacked model. I also worked on another research based project where I devised a proctoring system to detect malpractices during distant examination using computer vision, titled “Guided Vision Controlled Proctoring System for Distant Examination”. For this novel technique, which uses Body Tracking, Gaze Estimation, Emotion Recognition and Prohibited Object Detection, I have filed and published a patent in The Patent Office. These experiences have served as a great learning opportunity and reinforced my desire to further my knowledge acquisition through the Master’s program.
My undergrad me to implement the above-mentioned projects successfully. Subjects such as Machine Learning, Data Warehouse and Mining, Artificial Intelligence and Soft Computing: Introduced me to various concepts of artificial intelligence and machine learning such as various Supervised and Unsupervised Machine Learning algorithms, Neural Networks, Fuzzy set theory and data management and processing. Robotic Vision: Introduced me to Computer vision and helped me develop various projects using Computer Vision. Applied Mathematics, Discrete Structures and Graph Theory, Digital Signal Processing: Gave me the mathematical, statistical and analytical knowledge for understanding different machine learning algorithms, data processing techniques and optimization techniques. Advanced Data Structures, Design and Analysis of Algorithms: Gave me a complete understanding of various data structures and their uses, different algorithms like path finding, sorting etc. and space and time complexity. Software Engineering, Object Oriented Programming: Understood good practices of the software engineering process and understood the importance and 4 pillars of object oriented programming. Web Technology Lab, Enterprise Technology Lab, Minor Project Lab: Understood web and application development along with using databases , servers and GPUs.
To gain understanding of blockchain technologies, I worked on a research paper ‘Asset Tracking System using Blockchain’ which is awaiting publication in IEEE and presented at ASIANCON Conference. This paper presents a decentralized system to track status of an asset throughout the supply chain, making sure that the asset reaches the market tamper-proof and authentic. For this I used Solidity Smart Contracts, Ethereum Blockchain, Truffle Framework and Node.js. Additionally, I also worked on a decentralized Medical Record Dapp, to provide a safe and secure way of holding medical records of the users and share peer to peer with the concerned other users. The most important challenge was storing documents in a secure way but not in the blockchain as it would limit the size of the documents. Hence, we made use of IPFS (InterPlanetary File System) to store documents and make sharing possible in a secure way. These projects made me keen to gain a deeper understanding of the distributed systems through the graduate program.
To gain industry exposure, during my undergrad I took up an internship at J. P. Morgan Chase & Co. as a Software Developer Intern. Here I developed and deployed a Service Virtualization System for an ongoing project that simulates external services required for the development process. This was achieved by setting up mock APIs to eliminate the cost associated with usage of external APIs, using Wiremock with Moneta Boot and Oracle Database. Working here, I learned various efficient and cost-optimized methods of software development. Additionally, due to my experience of ML and blockchain, I also gave presentations to my team on ‘Getting started with Blockchain’, ‘Developing Machine Learning Projects’ and several others.
My professional experience also includes working at Modern Monk Pvt. Ltd. as a Flutter and Node.js Developer. Here, I Devised and deployed an application ‘Modern Monk’, a health management service consisting of dietician consultation, diet plans based on user requirements and training sessions using Flutter and Firebase. I also developed a backend API for dieticians to manage customer data and allocate diet plans using Node. Understanding the application requirements and developing the best suitable application. Learned how to structure unstructured data and serve it via a backend to the application. Understood the needs of the medical industry and practical applications of technology.
To contribute to the society using our Computer Science skills, I worked on a college funded project partnered with an NGO to make a mobile application, VRecycle which is currently live on Play Store. Using this app, the user could request solid waste pickup by specifying content of the waste, a picture and a suitable time slot, the collector could accept a pickup request and go pick up the waste at the specified time. Here we developed our own algorithm that chooses which collector to send the request to based on the location of the user, schedule of the user, the number of pending requests of the collector, user and collector’s distance etc.
Extracurricular activities:
Committee Work:
1) Training and Placement Coordinator, Sardar Patel Institute of Technology, from 2019. Devised training strategies, managed placement processes and Enterprise Resource Planning for the Training and Placement Cell.
2) Technical Head, TechRace (App Based Treasure Hunt Across Mumbai), Oculus (Techno-Cultural Festival) in the year 2019-2020. Developed an android and an iOS application for the participants to play the treasure hunt.
3) Public Relations Head, Computer Society of India (CSI), 2019-2020. Held various technical and cultural events, devised public relations strategies and handled social media accounts to increase the popularity. Also, organized a national level hackathon.
Sports:
4) Volleyball Team, Sardar Patel Institute of Technology, 2017-2019. Represented college at various volleyball tournaments.
Social Service:
5) Student Counsellor, Vishwas, 2019. Interacted and counselled freshers to help them acquaint themselves with the college.
6) Participated in various beach clean-up drives across Mumbai organized by Afroz Shah Foundation in the year 2018-2019.
Having worked on different areas of AI such as CV (video), audio analysis, and NLP, I realize that CV and NLP is the cornerstone of the AI industry transformation. Going forward, I would like to be part of the Artificial Intelligence research team of a tech giant company like Google, Amazon or Apple and utilize my knowledge gained from masters program to contribute to the ongoing pioneering discoveries in the field of Deep Learning and Reinforcement Learning while equipping myself with each and every bit of knowledge about the subject. In the long term, I aspire to establish a company that solves various social and corporate problems using Artificial Intelligence alongside researching to make pioneering discoveries in the field advancing it further.
Final Draft
I vividly remember how was I was completely awe-struck when Sunder Pichai introduced the AI abilities of Google Duplex in 2018, wherein the AI assistant handled a telephonic conversation to book a salon appointment independently. It demonstrated a “natural” flair in handling conversations with pauses and fillers such as hmm and huh. Who would have imagined that a few algorithms and commands could outpace human ingenuity and versatility one day – by making machines think, act, and speak like humans. The recent release of Google’s LaMDA, which can converse on almost everything, is a big step toward more human-like conversations. Impressed and intrigued by the vast potential of AI, I decided to embark on a journey to sharpen my skills and gain experience in such technologies by opting for an MS in computer science.
I gained a good insight into AI’s sentient capabilities in decoding human emotions while working on my 5th-semester project, “AI Therapist.” The goal was to recognize and analyze human emotions and, accordingly, recommend a suitable therapy to overcome the negative emotions of subjects. I trained CNN and BERT models for emotion recognition using image/video and text, respectively. Instead of converting the audio input to text and using the same BERT model, I converted audio signals into MFCCs, which became another CNN input. This project also featured a chatbot. After training the models, the designed prototype gave an accuracy of 89.92%.
The project’s outcome motivated me to understand NLP deeper through a self-initiated project: Exam Preparation Kit Using Natural Language Processing. The system allows students to speak out their answers, converts them into text, and punctuates and evaluates it to generate scores with detailed explanations. Based on the subject data, the system compiles questionnaires by factoring in students’ weak topics. The main challenge was evaluating students’ answers using sample answers. For this, I explored NLP concepts such as similarity measures and keyword extraction algorithms, to analyze and process the text better. I finally came across a sentence-BERT model that uses Siamese and triplet networks that fit our use-case perfectly and gave satisfactory results. The findings were documented in a research paper presented at the 12th ICCCNT 2021, which awaits publication in IEEE.
Alarmed by the rise in airplane crashes, I authored a paper titled “Airplane Crash Severity Prediction Using Machine Learning,” which was also presented at the 12th ICCCNT(full form)2021 and is awaiting publication in IEEE. For this, I developed several machine learning models, including a neural network and a stacked model using the stacking ensemble technique. These models could predict the severity of an airplane crash based on factors like location, weather, engine type, and flight purpose. Seeking to expand my research experience, I devised a proctoring system titled “Guided Vision Controlled Proctoring System for Distant Examination.” The objective was to detect malpractices during online exams employing novel techniques such as body tracking, gaze estimation, emotion recognition, and prohibited object detection. These invaluable experiences strengthened my desire to deepen my understanding of AI.
My undergraduate curriculum has laid a strong foundation to pursue a master’s program. I acquired an active comprehension of machine learning algorithms, neural networks, fuzzy set theory, and computer vision through the Machine Learning Lab, AI and Soft Computing, and Robotic Vision courses. Applied Mathematics, Discrete Structures and Graph Theory, and Digital Signal Processing introduced me to different machine learning algorithms and data processing techniques. I developed a deep understanding of data structures, algorithms, their uses and complexities, and code optimization through Advanced Data Structures, Design and Analysis of Algorithms, and Object-Oriented Programming. Database Management Systems, Web and Enterprise Technology Lab, and Software Engineering taught me how databases, frameworks, APIs, and libraries are integrated during full-stack development.
Keen on researching new technologies, I authored the research paper, “Asset Tracking System Using Blockchain,” published in IEEE Xplore. It highlights how a decentralized system could be devised and used to track product status through a given supply chain framework to ensure that products reach consumers without any tampering. I further supported my learning by developing Medical Record DApp, a safe and secure way of storing medical records. Since storing large documents on the blockchain is expensive and unfeasible, I integrated IPFS with the system to store and share documents securely. The project gave me hands-on exposure to Solidity smart contracts, Ethereum blockchain, Truffle, and Node.js in an industrial setting.
My transition into the professional environment was a valuable learning curve. The internship with JPMorgan Chase & Co. saw me working as part of its Asset and Wealth Management Department. Here, I witnessed how mission-critical applications relied on multiple services for their optimized development. It signaled an opportunity to develop and deploy a service virtualization system to simulate the external services required for the development process. I incorporated dynamic simulations through WireMock with Moneta Boot and Oracle database. This assignment improved my awareness of optimizing software development processes. I also delivered team presentations on topics such as “Getting Started with Blockchain” and “Developing Machine Learning Projects.”
Subsequently, working at Modern Monk Pvt. Ltd. as a Flutter and Node.js Developer, aligned with my fascination for deriving value from unstructured data. Here, I devised and deployed Modern Monk, a health management mobile application that could provide users with dietician consultations, diet plans, and exercise plans based on their needs, health, and wellbeing. Since operations were carried offline before the application was launched, user data was unstructured and incomplete. I processed and structured the data by running Python scripts and exported it to Cloud Firestore. I also developed an API for dieticians and trainers to connect with users.
Participating in extracurricular engagements helped me develop a better perspective toward life. I created a mobile application VRecycle, for my college solid waste management initiative. Using Flutter and Firebase, I created an interface that connected users and waste collectors. The application allowed users to book a waste pickup appointment at their desired location and time slot, which was then allocated to a suitable waste collector using a custom algorithm. I also organized various training sessions and placement drives as a Training and Placement Coordinator and conceived several technical and cultural events as the PR Head of the Computer Society of India. I regularly participated in beach clean-up drives held by the Afroz Shah Foundation and mentored freshman students under the Vishwas initiative.
The enriching exposure to AI constituents such as CV and NLP, inspired me to build a firm foundation in this domain. I believe a master’s program will enhance my ability to deliver more concrete and tangible solutions for addressing pressing real-world problems related to healthcare, education, etc. I envision serving pioneering technology organizations such as Google, Amazon, or Apple and contributing to their research in deep learning and reinforcement learning. In the long term, I aspire to establish a company that solves various social and corporate problems using AI alongside research to make pioneering discoveries in the field.
Whether it is decoding genetic markers or human communication, the application of disruptive technologies such as NLP and CV fascinate me. This fascination motivates me to pursue an MS in computer science with a specialization in AI from ____ University. I am specifically looking forward to courses on ___,___, and ___, to build upon my existing knowledge to redefine computational costs and processing time for complex operations such as processing unstructured data and comprehending languages. Prof. ____’s research on ____ appeals to me for its emphasis on ____. I am hopeful that the learning, research opportunities, and mentoring at ____ will enable me to design, build, and deliver predictive, intelligent, and sentient applications that can transform businesses and foster.
My intellectual abilities, problem-solving skills, and extracurricular contributions convince me of having the credentials and commitment to excel in the program. Given my affinity for knowledge sharing, I aspire to foster a culture of value sharing and mutual learning through various engagements at the university. I hope the admissions committee evaluates my application favorably and grants me the opportunity to be a part of the _____ class.
