I'm a highly motivated and results-oriented researcher with a strong background in Electrical Engineering and a passion for applying Machine Learning (ML) to solve real-world problems, particularly in healthcare. My journey began with a solid foundation in Electrical and Electronics Engineering (B.Tech, National Institute of Technology, India), followed by an M.Eng in Electrical and Computer Engineering at the University of Toronto (UofT).

My experience spans across various roles, from backend development to cutting-edge LLM research. Currently, I'm a Machine Learning Associate at the Vector Institute (UofT), where I actively contribute to finetuning, deploying, and implementing Retrieval-Augmented Generation (RAG) techniques for medical Large Language Models (LLMs).

My primary research focus lies in leveraging the power of LLMs to improve healthcare outcomes. I'm particularly interested in:

• Developing and deploying LLMs for medical applications: My work at the University of Toronto with Prof. Mark Chignell involves building "Clinical Panda," an LLM that generates evidence-based explanations for diagnoses. I'm also passionate about "PIIguardLLM" at Scribble Data, a project focused on enhancing data privacy with advanced LLM-based masking techniques.
• Multi-modal learning for the medical domain: At the Wang Lab (UofT), I'm exploring the potential of combining different modalities (e.g., text, images) to construct a robust multi-modal LLM specifically tailored for healthcare applications.
• Explainability and interpretability of deep learning models: Beyond healthcare, I'm also interested in developing explainable AI methods. My independent research project involved creating a novel neuro-inspired sparse visual explanation algorithm for Convolutional Neural Networks (CNNs).

My enthusiasm for AI safety principles fuels my dedication to developing LLMs that are not only effective but also responsible and trustworthy for real-world healthcare applications.

• 👨‍💻 All of my projects are available at website

• 📫 How to reach me [email protected]