I am currently in my final year pursuing a BS in Data Science and Applications from IIT Madras. Previously, I secured a position among the top 3 students in the Master of Computer Applications program at Pondicherry University. My academic journey spans physics, computer science, and data science, giving me a strong multidisciplinary foundation. During my academic tenure, I was awarded for a project in the Modern Application Development course at IIT Madras and earned the Swami Vivekananda Merit cum Means Scholarship for consistent excellence.
My research focuses on applied artificial intelligence, computer vision, and hybrid machine learning architectures. I have published three IEEE research papers. My work includes developing a machine learning-driven clinical risk prediction framework for mortality assessment in pediatric ICUs and designing a hybrid Quantum-inspired LSTM architecture to improve forecasting accuracy in volatile financial markets. Furthermore, I built a real-time computer vision pipeline for automated violence detection in public spaces, which received the Best Paper Award at IEEE ICRITO 2025.
Beyond academic research, I have engineered multiple high-impact AI systems. As a Lead AI Engineer at C-DAC, I developed real-time anomaly detection and suspicious behavioral profiling systems for high-stakes government examinations. During my internship at the Ministry of Electronics and Information Technology (Govt. of India), I designed an NLP-powered system for monitoring and classifying offensive content on social media, and conducted applied research on clinical risk prediction using real-world hospital data. I am also actively building independent platforms like QueryCortex AI, an agentic AI reasoning platform combining NL2SQL and vector search.
I have a keen interest in Generative AI, Agentic Workflows, and Computer Vision. I am looking forward to collaborating with research groups to expand my technical knowledge and contribute to innovative, large-scale solutions in AI and Machine Learning.
Generative AI & Agentic Systems Applied Machine Learning Natural Language Processing (NLP).Data Engineering & Advanced Backend Architectures Computer Vision
Prediction of stock price remains a challenging task because financial markets are dynamic and often non-linear, unpredictable patterns. To deal with such unpredictable data, the architecture that’s created must identify both time based trends and hidden patterns in data. In this study, we explore a hybrid modelling approach by combining both the classical deep learning techniques with quantum computing principles. Specifically, a Long Short-Term Memory (LSTM) is combined with a Quantum Neural Network (QNN) to understand the benefits of using quantum in time series forecasting. Rather than depending on a single quantum configuration, multiple feature maps and ansatz circuit designs were evaluated to identify the most suitable arrangement for the QLSTM model. Resultant framework blends ideas from both classical computing and quantum approaches to improve how we predict patterns over time in real-life scenarios.
The growing demand to upscale surveillance over public places and in online media provided one of the significant reasons for the need for violence detection in surveillance systems and in the moderation of online content. This paper proposes a real-time violence detection system that uses ResNet50V2 for violent activity detection with high precision, along with timestamping of incidents in video streams. It still has a confidence threshold at 0.6 for reliable detection with fewer false alarms. Extensive testing reveals high performance, such as achieved accuracy at 88 %, precision of 90 % for violence, and an F1score of 88%. The system thus assures better public security by ensuring timely responses and a holistic analysis of violent events.
In an Intensive Care Unit scenario, estimating the risk of a patient dying is useful for improving patient care and efficiently utilizing limited resources. Traditional scoring systems are available, but modern and promising methods from data-driven techniques have been shown to improve mortality prediction, specifically to Pediatric Intensive Care Units(PICUs). This research focuses on developing and implementing machine learning prediction techniques for mortality risk that uses a large dataset extracted from a single hospital database comprised of multiple tables using rigorous data preprocessing and cleaning work. The multiple tables in the large hospital database provide a collection of patient data that this research effectively organizes and integrates. The report includes patient demographics, admission details, and ICU stay details. Patient demographics, admission specifics, and information about ICU stays are among the elements of the hospital database that aid in the synthesis of a patient profile. This profile combines biometric information, admission details, measures associated with ICU stays, and whether the patient had an outcome of survival or survival outcome. In this paper, an approach is demonstrated to work with a huge single hospital patient dataset to merge classes and perform the feature extraction process with the use of various machine learning models for mortality risk prediction of the patients. This work contributes to the field of medicine, where disparate data is integrated, and advanced machine learning algorithms are used to improve mortality risk estimates and clinical decisions.
Hello, I am Shib Kumar Saraf , currently in my final year pursuing a BS in Data Science and Applications from IIT Madras , alongside holdin…
I am seeking collaborators in Computer Science and Artificial Intelligence, with interests in AI, machine learning, data mining, and applie…
Hello, I am Leena and currently doing master from USC in Computer Science (Artificial Intelligence), I have keen interest in research showc…