About

I’m Tanvir Ahmed Khan, a final-year undergraduate student in Electrical and Electronic Engineering at BUET, with a deep interest in machine learning, natural language processing, and computer vision. My research centers around improving the efficiency and reliability of large language and multimodal models, especially in medical domains where accuracy and interpretability are critical.

Currently, I’m working on:

🔬 Compression Techniques for LLMs: Exploring the effects of structured pruning, quantization, and layer-wise modifications to reduce the size and inference cost of LLMs without sacrificing performance.

🧠 Instruction-Tuned Medical Multimodal Models: Fine-tuning large vision-language models (e.g., LLaVA) on custom skin disease datasets to perform classification and instruction-following dialogue generation.

My goal is to contribute to the development of scalable, trustworthy AI systems, especially for high-stakes domains like healthcare. I’m actively preparing for Ph.D. applications (Fall 2026) and open to research collaborations.