projects
Linking saliency with modulations along the visual processing hierarchy
Oct 2022 - Present
University of Tübingen, Germany
As a research assisstant under Prof. Andreas Bartels, my focus is on analyzing a large fMRI dataset (the NSD dataset) to create “brain-derived” deep learning models for activity prediction. I am currently designing saliency metrics for stimuli and identifying correlations with retinotopic activity in visual, parietal, and prefrontal regions of the brain. As part of the project, I am also conducting a comparative analysis of high versus low-level model predictions for spatial modulations in the cortex, with a specific focus on model complexities.
Inferring neuronal computation from partial observations
Feb - Aug 2022
Allen Institute of Brain Sciences, Seattle
Under Prof. Kaspar Podgorski, I investigated methods for inferring a neuron’s computations from incomplete observations of its inputs and outputs. Using PyTorch, I simulated neurons as hierarchical linear-nonlinear models under various conditions, and utilized gradient descent methods to infer the parameters of the simulated neuron. This was part of my work for my master’s thesis
Investigating the physiology of aging across genetically diverse mouse population
Aug 2021 - Jan 2022
EPFL, Switzerland
Working under Prof. Johan Auwerx, I analyzed biological time-series data with machine learning and computational modeling. I developed a health index to summarize aging properties across genetically diverse mouse populations, and explored methods for analyzing time-series data across these populations.
Cloud Infrastructure Automation for Health Checks and Maintenance
** May - July 2021
ServiceNow, Hyderabad, India
As a software development intern in the cloud computing and development operations (DevOps) team, I worked on a proprietary cloud infrastructure platform for monitoring the health of cloud instances and load balancing. I developed tools for automating cloud instance health checks and maintenance.
Formal Language Modeling of Biological Networks
Aug - Dec 2020
Department of Computer Science, BITS Pilani
Working under Prof. Rajesh Kumar, I analyzed biological networks such as gene regulatory networks (GRNs) using formal language and automata theory. I focused on fault analysis of GRNs to understand the effects of induced drugs, and modeled stochasticity in GRNs using probabilistic boolean networks.
Machine Learning for Coronary Heart Disease Prediction
Aug - Dec 2020
Department of Biological Sciences, BITS Pilani
Working under Prof. Syamantak Majumder, I used machine learning and deep learning models to understand the biological parameters influencing heart diseases. I compared various models for their performance in predicting coronary heart disease using data from hospital patients.
Analysis of the neural patterns betweeen face and shape recognition
Neuromatch Academy
July 2020
The Neuromatch Academy was a summer school where we were trained on learning computational tools and apply them to real-life neuroscience research problems. Some of the topics which interested me the most were dynanimcal networks, bayesian statistics, and optimal control. As part of the group project, we worked on the Human Connectome Project (HCP) dataset, where we analysed the neural patterns between face and shape recognition in the brain.
Protein engineering with deep learning
RWTH Aachen University, Germany
May - July 2020
I worked as a research intern at the Institute of Biotechnology under Prof. Mehdi D. Davari. I worked on the prediction of biophysical properties of enzymes created through directed evoultion, a process mimicking natural selection. I experimented with various machine learning and deep learning models for choosing the best performing model for the required task, and analyzed high throughput data for protein mutation samples.
Machine learning for cancer prognosis using biomedical image data
Department of Biological Sciences, BITS Pilani
May - July 2020
I worked under Prof. Meghana Tare to analyze biomedical image samples with machine learning and deep learning models. I performed biomedical image segmentation with the help of convolutional neural networks, and classified the MR images using the 3-D voxel data.
Custom JupyterHub AWS instance with automated load balancing using Kubernetes
UST Global, Trivandrum, India
May - July 2019
I worked as a software development intern at the research center of the company, Infinity Labs. I worked on developing a proof of concept AWS JupyterHub cluster using Kubernetes for internal company resources to explore and prototype machine learning models without local installation and dependices. I also worked on automating the load management for each user’s instance, along with a custom login protocol based on an internal authentication method of the organization.