Neuromorphic Code Compiler. Tokenizes EEG brain-neural patterns and compiles neural intent into functional programming syntax (Node.js/Flutter) using deep learning representation architectures.

Context-Aware UI Adaptation for Zero-UI Environments. Implements spatial layout optimization, dynamic privacy masking, and eye-gaze tracking telemetry algorithms for next-generation AR/XR glass interfaces.

Biodegradable Digital Logic & Computational Matter. Simulates and models Biological Processing Units (BPUs) using DNA-based storage and organic protein logic circuits to eliminate global electronic waste (E-waste).

Decodes and digitizes molecular chemical structures into digital odor codes (similar to RGB for vision) to enable scent transmission, virtual reality integration, and volatile organic compound medical diagnostics.

Plant-to-Human Translation Model. Captures, processes, and translates plant bio-electrical action potentials into actionable, human-readable text notifications using deep learning to transform agriculture and botany.

TrustMed-RAG: A Verification-Driven Retrieval-Augmented Generation framework. Implements dual-stage verification loops to guarantee factual accuracy and eliminate hallucinations in AI-driven medical decision support systems.

"Robustness and Reliance: How Bangla–English Code-Mixing Affects Clinical Language Models and the Clinicians Who Use Them". Evaluates clinical LLM vulnerabilities and human-AI reliance under code-mixed medical corpora.

BanglaASR-Dial: An Automatic Speech Recognition framework for regional Bangla dialects. Fine-tunes Meta's Wav2Vec 2.0 architecture to accurately transcribe diverse, low-resource regional dialects into standardized Bangla text.

Implementation of a multi-modal data fusion framework for Early-Onset Alzheimer’s detection. Competently fuses retinal imaging, speech patterns, and genomic data to predict neurodegenerative diseases years before clinical symptoms manifest.