Projects
Primarily course projects alongside some engaging personal projects.
ML
Simple black box adversial attacks
Developed a robust pipeline for targeted and untargeted black-box adversarial attacks on machine learning models. Evaluated attacks' impact and vulnerability, providing valuable insights into machine learning models' susceptibility to adversarial examples. Created for the Spring'23 course - Statistical Methods in AI.
Empty parking lot detection
Developed an image processing algorithm to detect vehicle presence, count available parking spaces, and monitor vacancies in real-time, enhancing parking lot efficiency. Implemented solution successfully for both static images and video analysis. Created for the Monsoon'22 course - Digital Image Processing.
Movie Recommender
Built a comprehensive database from scratch and created a content-based recommendation system around it. Users select preferred films from a curated list of top 250 movies that align with their criteria. Employed vector transformation and cosine similarity on the selected movies to identify analogous movies. Designed with a focus on reducing user indecision, the system was complemented by a Flask-based website for seamless access.
Linguistics
SoBERT - Unsupervised text detoxification
Developed a pipeline to detoxify input by identifying, masking, and replacing toxic words, reducing the overall toxicity considering factors such as fluency, similarity to the original sentence, and perplexity. Created for the Monsoon'22 course - Advanced NLP.
Textual Coherence
Experimented with neural models (LSTM, GRU and RNNs) to measure textual coherence using parameters such as similarity on the GCDC and Wikipedia-CNN dataset. Created for the Spring '22 Course - Introduction to NLP
Covid Chatbot
Developed a user-centric bot addressing critical pandemic-related inquiries by offering real-time information on vaccinations, symptoms, and local hospital availability. Leveraged TensorFlow and Scikit-learn to enhance model accuracy and efficiency, while utilizing Tkinter to design an intuitive GUI for seamless interaction. Aided individuals in making informed decisions during the ongoing pandemic by providing reliable and up-to-date insights. Created for the Monsoon'21 course - Computational Linguistics II.
Twitter-Privacy Dashboard
Contributed to a project under the Digital Identity Research Initiative (DIRI) at the Indian School of Business (ISB) Hyderabad. Developed a dynamic dashboard for sentiment analysis of tweets, allowing analysis by company or industry. The project focused on observing trends in public sentiment related to data privacy violations, while also facilitating comparisons of company popularity based on retweets and likes. This initiative demonstrated proficiency in data analysis and visualization for impactful research.
Twitter Analysis - Lingua Franca of India
Developed a comprehensive database by curating tweets and subsequently designed an informative dashboard featuring data visualizations. This project aimed to analyze public sentiment and opinions on India's National Language or Lingua Franca. Demonstrated proficiency in data collection, analysis, and visualization techniques, contributing to a deeper understanding of societal perspectives. Created for the Monsoon'21 course - Language and Society.
Clickbait identification
Conducted an in-depth analysis differentiating between Clickbait and Non-Clickbait titles, employing SVMs, Logistic Regression, and XGBoost classifiers for accurate classification. Created for the Spring'21 course - Computational Linguistics 1.
Development
Vinci - Poster generator
Contributed to a project aimed at minimizing the DIY effort in photo editing on well-known platforms like Adobe Photoshop and Canva. Successfully achieved a reduction to 15-20% DIY effort through the creation of an event-specific poster generator, combined with an intuitive user customization interface. Developed the entire user interface from scratch, encompassing the event-specific poster generator and the intuitive customization tools, to enhance design efficiency and user experience in photo edit.
Healthcare Knowledge Graph
Developed a healthcare knowledge graph using Neo4j and MERN stack for clinical decision-making and drug discovery.Created a user platform with secure authentication and dataset upload, enabling personalized knowledge-graph access. Created for the Spring'23 course - Data Foundation Systems.
IIIT-Canteen portal
Engineered a MERN-based web application facilitating food item listing and transactions, catering to vendors and buyers. Designed features including real-time store open availability, buying and selling options, and favoriting. Leveraged Docker for seamless deployment, ensuring efficient user experience and streamlined operations. Created for the Spring'22 course - Design and Analysis of Software Systems.
Clash of Clans
Developed a Python3-based terminal 2D game reminiscent of Clash of Clans, featuring strategic control of a king or queen character. Implemented dynamic movement and interactive gameplay mechanics involving destruction of structures and combat against defenses, showcasing a strong blend of coding skills and game design Created for the Spring'22 course - Design and Analysis of Software Systems.
Checkers
Designed and implemented a terminal-based multiplayer Checkers (English Draughts) game in C, incorporating advanced functionalities such as displaying possible moves, review options, and undo capabilities. Created for the Spring'21 course - Data Structures and Algorithms.
C-Shell
Developed a fully functional Linux shell in C, encompassing a wide range of Linux shell functionalities, including basic commands, process control, redirection, tab auto-completion, piping, signal handling, and advanced parsing using semicolons and ampersands. Successfully incorporated foreground and background process execution capabilities, showcasing advanced programming skills in system development. Created for the Monsoon'22 course - Operating Systems and Networks.
Others
Gamification in GPay
Conducted an in-depth analysis of gamification principles through extensive research, drawing correlations between these principles and GPay's methods. Executed a comprehensive case study on GPay's gamification strategies, assessing the impact of introduced games on user engagement. Gained insights into user perspectives by both exploring the app and engaging with GPay and non-GPay users, enhancing understanding of effective gamification strategies. Created for the Monsoon'22 course - Social Science Perspective on HCI.
To Focus Attention - think on your feet not your seat
Explored the intricate relationship between body posture and cognitive processes, particularly attention. Designed and conducted experiments to investigate cognitive changes related to maintaining postural stability. Addressed the contemporary concern of studying in various postures, revealing insights into the impact on attention and productivity, which can guide students in optimizing their learning environment for enhanced performance. Created for the Monsoon'22 course - Introduction to Psychology.
Algorithm - Mini games
Designed and developed an educational website featuring interactive mini-games and puzzles to elucidate complex Algorithms, including greedy, dynamic programming, and divide and conquer. Empowered users to grasp algorithmic concepts through engaging hands-on experiences, fostering a deeper understanding of fundamental computer science principles. Created for the Monsoon'21 course - Algorithm Analysis and Design.