At AORVIS, I developed and integrated AI/ML solutions, software applications, and data-processing workflows using Python. I contributed to geospatial data processing, visualization, and domain-specific modelling.
At Xeta Labs, I was an ML Intern working on geophysical data modelling with CNNs, Faster R-CNN, and PyGIMLi. I annotated datasets with Roboflow and analyzed magnetometer and SRT data to estimate depth.
In a collaboration between University of Calcutta and University of Copenhagen, I worked on fine-tuning Large Language Models and optimizing hyperparameter reduction techniques. I used Hugging Face and Colab Pro notebook for experiments supporting natural language processing tasks.
I developed a machine-learning recommendation system for suitable crops and fertilizers using soil nutrients, pH, temperature, and rainfall. It uses CatBoost, Gaussian SVM, and MLP models, with Android and Streamlit applications for decision support.
