A passionate Machine Learning Researcher with experience in working on machine learning tasks involving multiple types of data including vision-based data, audio data, numeric as well as text-based data. I currently work as a Machine Learning Researcher with the BROAD Insititute of MIT and Harvard in Boston and have previously completed an ML research internship with the Massachusetts General Hospital (Harvard Medical School) where I worked on denoising MRI scans. In my present role, I have developed and trained 5 supervised cell-type classification models that accurately classify cell types, make use of the cell-type ontology structure through soft-constraints in model training and have evaluated these models using a novel hop-based F1 scoring technique that takes hierarchy of cell types from cell ontology into consideration at each hop level. Currently, I am benchmarking the results from my models against the leading cell classification techniques that make use of scRNA data and am editing a pre-print draft to present my work. When pursuing my MS degree at Northeastern University, I also worked as a Graduate Student Researcher working on generating image-based datasets to identify types (Fonts) and associating cognitive properties to these Types.
My repositories include code samples, papers explaining my work as well as some interesting research paper reviews!

• 📫 How to reach me nimishmagre@gmail.com

• 📄 Know about my experiences: Resume

- 📜 My first Arxiv paper: https://arxiv.org/abs/2202.08112>
- 📢 Some datasets on Kaggle: https://www.kaggle.com/datasets/nimishmagre/tmnist-glyphs-1812-characters>
- 🎶 Music samples generated through an LSTM based network: https://soundcloud.com/user76377803/sets/lstm-based-music-generation?si=4b10df90b2a241ed9897f801c7002634&utm_source=clipboard&utm_medium=text&utm_campaign=social_sharing>