Research & Open Source

Open-source contribution and reproducible science are extremely important. I maintain several personal research projects focusing on graph neural networks, link prediction, and network visualization.

Relevant Open-Source Projects

fm4tlp A foundational project offering transfer learning pipelines designed specifically for temporal link prediction in dynamic graphs. This research enhances models’ ability to predict evolving network connections over time.

ILP (Inductive Link Prediction) This work implements novel inductive approaches to predict edges for unobserved queries and temporal nodes. It addresses the challenge of making network link predictions without relying entirely on preexisting structural topology.

AI-Bind Implementation An AI-driven pipeline assisting in protein engineering workflows, targeting protein-ligand interactions. It uses deep learning models alongside interpretable active binding-site identification to accelerate therapeutic discovery.

MultiViz A specialized plugin developed for Gephi to provide scalable and intuitive visualization of multilayer and complex networks, allowing users to efficiently analyze interacting graphs.

Feel free to browse through my GitHub for a comprehensive look at the various libraries and data scripts I’ve built.