Namaste, Hello and Bonjour! I am Shashank Kumar Roy, a researcher (Ingénieur de recherche) working in Lab-STICC under Prof. Ronan Fablet at the Institut Mines-Télécom (IMT), Atlantique, Technopole-Brest, France.
I am currently working with 4DVarNet – a deeplearnig based data assimilation algorithms for numerical weather prediction. Broadly, I lay my interest in the rapidly evolving field of SCIML (Scientific Machine Leanrning) where we leverage deep leanring & data to solve problems in ocean system modeling, reconstruction and forecasting. We focus on improving the predictive performace of dynamical models that arise in numerical weather prediction problems. My curiosity lies in climate and geo-spatial data and developing their stable and robust data-driven models. My days are spent writing and fixing codes in Python using mostly in Pytorch and Jax.
Prior to this position, I was a PhD research scholar at International Center for Theoretical Sciences - Tata Institute of Fundamental Research, Bangalore, India, where I also finished my Masters in Physics. My PhD thesis titled,‘A study of dynamical instability and filter stabilit using ensemble Kalman filter’ is now available here on ICTS Thesis arxiv. In the thesis, I worked with Ensemble Kalman Filters, a class of monte-carlo based state estimation algorithm based on kalman filter that implements approximate bayesian posterior computation using model estimates and real data sequentially over time in an online approach. I completed my doctoral work under the guidance of my advisor Prof. Amit Apte & Prof. Samriddhi Sankar Ray at ICTS.
Before that I completed my Bachelors in Physics from Ramjas College, North Campus, Univeristy of Delhi, India.
Feel free to contact me if you would like to drop a line or you have a similar idea to discuss via e-mail shashankroy1997@gmail.com, shashank.roy@imt-atlantique.fr or linkedin