PhD Research Scholar (SRF) at IIT Kharagpur, developing reproducible methods for understanding hydroclimatic extremes and translating environmental data into decision-ready evidence.
I study the interactions among rainfall, antecedent wetness, terrain, and landslide occurrence in complex mountain environments. My work combines hydrometeorological analysis, spatial statistics, threshold modelling, and responsible AI to support reproducible science and practical early-warning applications.
- Hydroclimate extremes: rainfall variability, antecedent moisture, and compound hazards
- Landslide risk: rainfall thresholds, susceptibility controls, and early-warning diagnostics
- Environmental data science: geospatial analysis, statistical modelling, and reproducible workflows
- Responsible AI: human-supervised decision support, retrieval systems, and privacy-aware automation
| Project | What it delivers | Core tools |
|---|---|---|
| Moisture-aware rainfall thresholds | Reproducible analysis of rainfall, antecedent precipitation, root-zone saturation, and landslide thresholds across the Northeastern Himalaya | MATLAB, Python, statistics |
| ClimateCareer-Agent | Privacy-conscious, human-in-the-loop AI workflow for evidence-backed climate-career decisions | Python, LangGraph, Streamlit |
| Hydro-Meteorological AI Research Assistant | Retrieval-augmented exploration of hydro-meteorological research literature | Python, FAISS, LangChain |
Scientific computing: MATLAB · Python · R
Geospatial analysis: QGIS · spatial statistics · terrain and raster analysis
Data and modelling: pandas · NumPy · statistical learning · uncertainty analysis
Applied AI: retrieval-augmented generation · LangGraph · Streamlit · human-in-the-loop systems
I aim to make research code understandable beyond the original analysis. My public projects prioritize explicit assumptions, traceable methods, documented data requirements, runnable examples, and clear boundaries around privacy and responsible use.
I welcome conversations about hydroclimate extremes, landslide early-warning systems, environmental data science, and carefully governed AI for research. The best starting points are LinkedIn or a focused issue in the relevant repository.