From Algorithms to Applications: My Research in Machine Learning
Date: 26 April 2025
Machine learning has rapidly transitioned from a theoretical pursuit to a critical driver of innovation across industries. My research journey reflects this evolution - beginning with an interest in core algorithms such as supervised learning, unsupervised clustering, and reinforcement learning, and evolving into the development of applied solutions that tackle real-world challenges.
In the early stages, my focus was on understanding and implementing foundational algorithms — decision trees, support vector machines, and neural networks — and analysing their behaviour through metrics like accuracy, precision, recall, and F1-score. This foundational knowledge enabled me to compare models, optimize hyper parameters, and address overfitting and under fitting through regularization and cross-validation techniques.
As my expertise deepened, I shifted toward applying these algorithms to domain-specific problems such as: Finance (e.g., credit scoring and fraud detection), Natural Language Processing (e.g., sentiment analysis and chatbots), and Computer Vision (e.g., object detection and facial recognition).
A significant part of my research involved building end-to-end pipelines - from data pre-processing and feature engineering to model deployment and monitoring. I also explored the ethical implications of AI systems, emphasizing the importance of fairness, accountability, and transparency in ML applications.
Additionally, my work included performance benchmarking, scalability tests using distributed systems (like Apache Spark), and experimenting with advanced architectures such as convolutional and recurrent neural networks, and more recently, transformers.
Research Empowerment Talk
Date: 28 January 2025
The journey of a Ph.D. begins with a vision—a dream of making a meaningful contribution to a field of study. It starts with a compelling question that sparks curiosity and drives innovation. Defining a clear and impactful research objective, aligning goals with personal interests and societal needs, and conducting thorough background research are essential steps. The path of exploration is marked by continuous learning, resilience, and collaboration. Embracing setbacks as opportunities for growth, seeking mentorship, and leveraging technology enhance research efficiency. Encouraging faculty mentorship, organizing brainstorming sessions, and promoting interdisciplinary collaboration can further foster innovation and tackle complex challenges effectively.