I am a fifth-year Ph.D. candidate in Informatics at Penn State's College of IST, advised by Dr. Syed Billah at the a11y lab. I hold a Bachelor's in Computer Science and Engineering from Bangladesh University of Engineering and Technology (BUET) and a Master's in Informatics from Penn State. My research connects Accessibility, HCI, Computer Vision, and ML Systems — with a focus on the gap between what works in evaluation and what holds up in the real world. Please check out my Portfolio for more details on my research.

On the HCI side, I build and evaluate interaction tools for blind and low-vision users across desktop, mobile, and VR. The common thread is treating accessibility barriers as engineering problems — ones that can be measured, modeled, and systematically solved. My research spans rethinking input device design to reduce navigation overhead, optimizing screen layout to lower cognitive load, and developing probabilistic metrics that capture whether an accessibility improvement actually worked.

On the Vision and ML Systems side, I focus on the mismatch between how models are evaluated and what real deployment actually demands. That's led me to build dataset auditing tools and heatmap-based debugging infrastructure for vision-language models, develop improved training objectives for temporally consistent video segmentation, and design a telemetry-driven narration pipeline that matches strong multimodal baselines without ever processing a video frame.

Beyond research, I bring 3+ years of industry engineering experience in distributed and backend systems, and I teach Distributed Systems at Penn State — two perspectives that keep me grounded in the gap between research and production tradeoffs. My first-author work has appeared at CHI, UIST, ASSETS, IMWUT, and DIS, with Best Paper Honorable Mentions at UIST'24 and DIS'25.