Abin Shakya
Postdoctoral Research Scientist
Biography
Abin Shakya is a Postdoctoral Research Scientist at Columbia University working in AI for Science, scientific machine learning, and high-performance computing. He completed his PhD in Computer Science at Louisiana State University, where his research focused on applying artificial intelligence to scientific challenges across materials science, Earth science, chemistry, and computational physics. His work includes machine learning for molecular simulations, graph neural networks, and LLM-assisted scientific information extraction and knowledge discovery.
Research Interests
Machine Learning for Scientific Discovery
Developing AI and machine learning methods to accelerate discovery and understanding across disciplines of science and engineering.
AI Agents for Scientific Automation
Building autonomous AI agents that integrate reasoning, planning, and tool use to automate complex scientific workflows such as molecular dynamics simulations and data-driven experimentation.
LLMs for Scientific Knowledge and Reasoning
Building and fine-tuning LLMs for scientific reasoning, automated information extraction, and dataset construction to accelerate data-driven research.
Latest News
Started new position: Postdoctoral Research Scientist
Working on AI for Science, scientific machine learning, and computational workflows for mineral physics and materials science.
Completed PhD in Computer Science
Successfully completed my PhD in Computer Science at Louisiana State University with research focused on AI-driven scientific computing, molecular dynamics simulations, and machine learning for materials science.
Finalist — ESA Antlion Pit Competition 2025
Our Smart Orchard Ant Monitoring team has been selected as a finalist for the Antlion Pit Competition to be held during Entomology 2025 in Portland, Oregon. The project uses AI-based vision systems and IoT sensors for real-time detection of ants and mealybugs in citrus orchards.
Learn MoreAdvanced to Ph.D. Candidacy
Successfully defended my dissertation proposal and officially advanced to doctoral candidacy.
Paper Accepted at KDIR 2025
Our work on "Weakly Supervised Graph Neural Networks for Scalable 3D Phase Segmentation in Molecular Dynamics Simulations" has been accepted!
Selected for KDD 2025 PhD Consortium
Selected to present my work “Density-Aware Phase Segmentation in Sparse 3D Atomistic Data” at the KDD 2025 PhD Consortium.
Shell Summer Research Fellowship
Selected as a recipient of the Shell Summer Research Fellowship to support my summer research on AI for molecular dynamics simulations.
Paper Accepted in Scientific Reports
Our paper “Insights into Core-Mantle Differentiation from Bulk Earth Melt Simulations” was accepted for publication in Scientific Reports, presenting ML-based molecular dynamics simulations that reveal new insights into elemental partitioning and core–mantle differentiation in bulk Earth melts.
Learn MorePresented our work at ANPA Conference 2023
Presented our research on ML-based molecular dynamics simulations for studying metal–silicate differentiation in bulk Earth melts at the ANPA Conference 2023. This work was later expanded and published in Nature Scientific Reports.
Learn More