Shahir Abdullah

Pursuing thesis based Master's in CSE at United International University (UIU), Bangladesh;
with research interests in AI/ML, Intelligent computing, Human-Computer Interaction, agents, adaptive UI systems, distributed systems, and computational optimization.
Completed my B.Sc. in CSE from Military Institute of Science & Technology, Dhaka, in 2021.
During my undergraduate studies, I developed an assistive Bangla book-reader system for visually impaired users, combining Android, Arduino-based hardware, computer vision, machine learning, and text-to-speech technologies.
Currently, working as a Software Engineer at Ajentica.ai, where I lead a small team involved in LLM-based desktop UI automation, exploring structured UI representations, workflow-aware UI grounding, graph-based UI optimization, and adaptive automation for changing software interfaces.

Email  /  CV  /  Bio  /  Scholar  /  ResearchGate  /  Github

Research Interests

I'm interested in AI/ML, Optimization, HCI, Deep learning, Distributed systems, Cloud computing, UI Automation, Generative AI.

Publications & Research Projects

A lightweight, smartphone-based safety-check system. At registration a soft identifier (a phone number) is recorded; after the physician writes a prescription, its image is captured, the brand names are resolved to active ingredients using national drug references, and the ingredients are matched against the patient's recorded severe reaction history. The system is retrieval-based rather than predictive, and is silent by default, raising a flag only for high-risk matches.

A two-dimensional model in which each distributed database configuration is represented as a point (H, V), with H denoting node count and V a vector of resources.

A desktop application that recognizes four basic hand-drawn geometric shapes, circle, square, rectangle, and triangle using a compact Convolutional Neural Network (CNN).

Graduate, Undergraduate and Personal Projects

Android app that takes pictures of book pages, converts image to text using EfficientNetB3 model, and converts to speech

System Analysis, Design & Development course project featuring emergency help, restricted internet access, video calls, messaging, file sharing, and discussion sessions with mental health experts.

Comprehensive Java tool implementing lossless compression (RLE, DPCM, predictive encoding, entropy encoding, Huffman encoding), lossy compression (DCT, wavelet transform), and steganography watermarking for image encoding.

An open-source Python toolkit implementing sixteen classical digital image processing techniques spanning intensity transformations, histogram-based processing, spatial (neighborhood) filtering, and edge detection. Each technique is implemented as a self-contained Jupyter notebook using NumPy, OpenCV, and Matplotlib, with several notebooks pairing a from-scratch, pixel-loop implementation against OpenCV's built-in equivalent for validation.

Miscellaneous

Template stolen from Jon Barron's Site.