Machine Learning Researcher
(Graph Neural Nets, Biomedical AI, AI Agents, Human-centric AI & NLP)
Kaggle Grandmaster | Explorer | Looking for research opportunities
About :
- An aspiring AI researcher and engineering student, exploring Graph Neural Networks (GNNs) in Bio-Medical AI, mainly focusing on neuro, therapeutic, and molecular ML domains (AI4Science). Along with GNN, my other research interests include AI for Science, Human-Centered AI (HCI, HAI) with NLP for interdisciplinary works.
- I am looking forward to pursue a PhD in Spring/Fall 2026 to continue research and looking for potential options.
- Currently, I'm working with Riashat Islam at Mila Quebec on Agentic AI and Generative AI, and computational biology - molecular ML. Previously, I worked with Prof. Dong-Kyu Chae at Hanyang University for 2 years on GNNs, Medical AI, and HCI-HAI. Additionally, I founded CIOL to mentor young researchers and bridge the gap between Industrial Engineering and AI. Here, I collaborate with Prof. Mahathir M Bappy (Louisiana State Uni.) and Prof. Manjurul Ahsan (Uni. of Oklahoma) on GNNs, Digital Twins, AI4Science, PINNs, and Medical AI applications; and guide young researchers. I am also a Contributor in Cohere for AI Aommunity and regularly work on Cohere projects. I work as Research Scientist at HerWILL, too. I'm also the 3rd Kaggle Grandmaster of BD.
- My works has been published in prestigious venues such as ICLR, WWW, COLING, DASFAA, CSCW, ACCV'24, Workshops of NeurIPS, AAAI, ICML, ACL and CHI, with ongoing reviews in some others.
- Outside research, I have work experience in AI-integrated IT Automation, Project - Product Management and Analytics roles.
- Passionate about learning new things, sharing my knowledge, improving myself regularly, experimenting with acquired skills and challenging my capabilities. Building all-in-one free AI/ML resources collection here.
Research :
- 🧬 Exploring Computational Biology and Biomedical AI Applications: Working on Computational Molecular Biology, Bioinformatics, and Drug Discovery, focusing on molecular property prediction, protein discovery, and binder design. Experienced in de novo protein generation, RL/energy-guided modeling with GNNs, Flow Matching, GFlowNets, and Diffusion models. Exploring Digital Twins, clinical reasoning and Agentic LLMs for clinical/biomedical applications
- 💠 Understanding and Applying Graph Neural Networks (GNN): Engaged in Geometric Machine Learning, focusing on GNN theories and applications, with an emphasis on biomedical AI applications as described above. Additionally, applying GNNs to industrial engineering, specifically in supply chain optimization and manufacturing. Interested in the application of knowledge graphs (extending GNNs) across diverse systems and domains.
- 🧑💻 Interdisciplinary Research on Humans, AI, and Language: I’m working on RL and reasoning in LLMs, focusing on self-verification, uncertainty estimation, and agentic decision-making in generative models, with AI4Good applications. My recent work develops trustworthy AI reasoning by integrating RAG, reward-based RL fine-tuning, Digital Twins, and structured reasoning to enhance model reliability. Additionally, in AI for Good, I’ve worked on multilinguality, accessibility, fairness, human factors, and cultural values in Generative AI and LLMs.
- Languages: Python (Advanced), C, C++, MATLAB, R, SQL
- DS & ML Tools (Python): NumPy, Pandas, Matplotlib, Seaborn, Scikit-learn, TensorFlow, PyTorch, LangChain, VLLM, Pydantic
- Data Science Techniques: EDA, Experiment Design, Hypothesis Testing, Sampling, and Data-Driven Decision Making
- Machine Learning Techniques: Statistical ML Methods, Deep Learning, NLP, Computer Vision, Graph Neural Networks (GNNs), GFlowNets, Flow Matching, Diffusion Models, RL and Reasoning in LLMs, Self-Verification, Uncertainty, Agentic Decision-Making, AI Reasoning, RAG, and Reward-Based RL Fine-Tuning
- Biomedical AI and Clinical Applications: Molecular Properties, Binder Design, Molecular Interaction, De Novo Protein Design, GNNs, RL/Energy-Guided Modeling, Generative Modeling with Flow Matching and Graph Diffusion, Reward-Based Generative AI, Agentic LLMs, Knowledge Graphs, AI-Based Drug Discovery and Genomics
- Interdisciplinary AI Research: AI for Good, Multilinguality, Accessibility, Fairness, Human Factors, Local & Cultural Values
- Others: GitHub, Collaborative Tools (AMs, VS Code, Azure, AnyScale, Replit, Colab, Kaggle), Parallel & Distributed Computing
