About Me
I’m a reseaerch technologist at Northwestern University. As part of my research I have worked on signal processing and machine learning algorithms for brain computer interfaces and neuroscience. My interests have been in learning representations in the brain and building decoders that help in translating neural signals to behavior. In addition to my interest in neuroscience, I am always on the lookout for interesting projects to work on. I am currently intersted in exploring machine learning tools used on healthcare data and am open to collaborating on projects, feel free to contact me via email or linkedin.
Projects
Book recommendations
Analyzed conventional and deep learning algorithms for recommending books to different users based on their rating for some books in the database. Applied Clustering techniques to identify similar users according to their ratings.
Sentiment Analysis on twitter dataset
Identified the sentiment of tweets denoted as +1 for happy and -1 for sad. This involved preprocessing the tweets to remove certain stop words and characters. Further, apply feature engineering techniques to represent the data appropriately for classification.
Identification of spam email
Built a classifier to identify email that might be spam. This involved first training on data that was already labelled as spam/not spam. Observed how changing certain parameters in the classifier affect the accuracy. Performed extensive search for hyper-parameter optimization.
Experience
Slutzky Neuroprosthetics Lab
Research technologist
July 2019 - Present
- Myoelectric computer interface for stroke affected patients: Conducting R01 study for stroke affected patients with upper limb paralysis using Myoelectric signals. Leading a pilot study that combines MCI with a sleep component to compliment rehabilitation.
- Decoding produced speech using ECoG signals: Designing methods to build speech decoder for patients with aphasia. Investigating deep learning methods to decode intended speech.
Kording Lab
Graduate Research assistsnt
Feb 2018 - May 2019
- Ensemble learning for classification of brain signals: Applied feature engineering, signal processing and machine learning techniques on EEG signals to decode motor imagination. Achieved significant accuracy using ensemble classification techniques like XGBoost and Random Forest
- Computational models of Biologically plausible neural networks: Helped design an algorithm to address the weight transport problem in neural networks. Successfully tested algorithm in a variety of neural network architectures. Extensively experimented with deep learning architectures using Tensorflow to address this problem.
Education
University of Pennsylvania
Msc - Electrical Engineering
2017 - 2019
During my Masters, I actively sought coursework and projects in the field of machine learning, systems engineering and neuroscience. I worked on research that focused on learning representations and brain machine interfaces.
Vellore Institute of Technology
Bachelor of Technology - Electrical and Electronics Engineering
2013 - 2017
My undergraduate education helped engender my fascination for signal processing and machine learning algortihms. I participated in hackathons and competitions that helped me appreciate the strength of interdisciplinary backgrounds for the success of a project.
Skills
- Programming languages: Python,SQL, MATLAB, C++
- Tools: Numpy, NLTK, Tensorflow, Keras, Pandas, sklearn, scipy
- Technical: Machine learning, Deep learning, Data cleaning, Data Analysis