Algorithms with a path to impact
Research
I study graph learning, AI systems, and optimization methods that balance competing objectives in scientific and engineering applications.
Graph learning at scale
How much of a graph do we need to learn well? I design sampling and sparsification algorithms that retain useful structural and attribute information while reducing computational and memory costs.
My work spans supervised edge selection, support graph theory, and reusable edge-disjoint decompositions, with attention to both homophilic and heterophilic graphs.
Learning for system design
I use graph representations and learning-based optimization to connect computation structure with hardware and software design decisions.
Current directions include differentiable hardware/software partitioning and scheduling, reinforcement learning for joint optimization, and graph-based circuit equivalence checking.
AI for cybersecurity
I develop language-model and knowledge-graph methods that connect vulnerabilities, weaknesses, attack patterns, and adversary behavior.
The aim is to make fragmented security knowledge easier to analyze at scale, from hierarchical vulnerability classification to threat intelligence and risk prioritization.
Multimodal & agentic AI
I build AI systems that extract, organize, and retrieve information from large collections of scientific and regulatory documents.
Projects include the NEPATEC corpus, multimodal metadata extraction, map-based georeferencing, and retrieval-augmented assistants for environmental permitting.
Multi-objective & many-objective optimization
I design evolutionary algorithms for problems with multiple competing objectives, with a particular focus on many-objective optimization and the balance between convergence toward the Pareto front and diversity among solutions.
My work includes F-DEA, which combines adaptive fuzzy dominance with reference-point guidance, and a fuzzy-dominance approach with bidirectional bias developed in my undergraduate thesis. I also apply many-objective methods to computing-cluster performance and scheduling, including hybrid greedy and nondominated sorting approaches. This research formed the basis of my M.Sc. and B.Sc. work at BUET, advised by Md. Monirul Islam.
Research in practice
Explore the projects.
Current work, public code, and interdisciplinary applications.