Work Experience

Cobra Groeninzicht, Vianen

Graduate Trainee: LiDAR data segmentation using Deep Learning
Duration: Sep 2023 - present
What do I do?
I work with deep learning models such as PointNet and PointNet++ for LiDAR data segmentation. I make use of Python for training the model and Cloud Compare for preparing training dataset.

University of Twente, Enschede

Geospatial Data Scientist: Critical water shortages and ground water gamble
Duration: Jul 2023 - present
What do I do?
I work with large‑scale hydro‑meterological datasets and Indian census datasets. I am responsible for creating data pipelines, including data collection using APIs, processing and generating climate indices, and integrating interdisciplinary datasets. These pipelines are mostly created in Python and R. One of my notable accomplishments in this role was the successful creation of a multi‑scale monthly drought index dataset for pan‑India, spanning over the past 43 years. Our project’s achievements were recognized when we secured funding from the Climate Centre Seed Fund to support another project within the university.

Researcher: On the move: Fine‑scale climate migration modelling in India
Duration: Mar 2023 - Jun 2023
What I did?
I worked with handling Big-Geo data for creating climate indices and integrating these dataset with social science dataset. I made use of Python and R programming for computing climate indices using libraries such as Xarray, GeoParquet, Dask for parallel computing. One of the key challenges was handling such a large scale dataset.

Researcher: Multi‑spatial‑temporal characterisation of changing peri‑urban areas of Chennai, India.
Duration: Dec 2022 - Feb 2023
What I did?
I performed a spatio-temporal classification of Landuse Landcover using machine learning. I used these maps to calculate spatio-temporal metrics using Python. I was able to identify areas with specific spatio-temporal charactristics.

University of Padova, Italy

Research Intern: Application of improved U‑Net models to detect landslides using spectral and topographic information
Duration: Jan 2022 - Apr 2022
What I did?
I utilized state-of-art U-Net models to detect landslides. My tasks were to prepare training dataset, train and optimize models and finally test and evaluate the model.