Selected Projects & Pedagogy
A curation of technical explorations and educational contributions, bridging the gap between theoretical research and pedagogical practice.
Research Projects
04 Featured ArtifactsGraph Tabular Learners
Investigating how to integrate traditional tabular learners with graph structured representation of datasets.
Dataset Condensation for Machine Learned Interatomic Potentials
Generalizing graph dataset condensation beyond simple graphs to geometric graphs with SO(3) rotational equivariance.
Causal Graph Learning for Air Pollution & Health
Developing methods to learn causal graphs from spatio-temporal epidemiological data, with a focus on identifying the health impacts of air pollution in Delhi. Combines causal discovery, time-series modeling, and environmental data to uncover interpretable relationships between pollutants and disease outcomes.
Graph Tokenization for LLM-Augmented Learning
Designing ways to encode graphs as compact, information-rich token sequences aligned with how LLMs process text. This enables LLMs to reason over structured data and inject knowledge from scientific literature and web-scale corpora—boosting performance on tasks where labeled graph data is scarce.
Read Related ArticleTeaching & Pedagogy
ExperienceData Mining
COL761: Undergraduate Core Requirement
Data Mining
COL761: Undergraduate Core Requirement
Graph Foundation Models
COL870: Undergraduate Elective
Data Mining
COL761: Undergraduate Core Requirement
Data Mining
COL761: Undergraduate Core Requirement
Graph Foundation Models
COL870: Undergraduate Elective
Data Mining
COL761: Undergraduate Core Requirement
Deep Learning
COL870: Undergraduate Elective
Machine Intelligence & Learning
ELL409: Undergraduate Core Requirement
Philosophy
"The cosmos is within us. We are made of star-stuff. We are a way for the universe to know itself." — Carl Sagan