May 2021 – Jul 2022 · 1 yr 2 mos · United States

Sep 2015 – May 2021 · 5 yrs 8 mos

Jun 2014 – Aug 2014 · 2 mos · Greater Minneapolis-St. Paul Area

• Analyzed Big Data (2 TB) to understand effects of Vagal Nerve Stimulation (VNS) on Heart Rate. A patent application was filed by Boston Scientific on this work in Oct 2014 on which I am listed as an inventor.
• Responsibilities included:
• Working closely with a team that includes physiologists, statisticians, scientists, engineers and clinicians
• Learning quickly the physics behind VNS from colleagues and literature survey
• Building statistical models that can capture most effects of the therapy on Heart Rate
• Writing computationally efficient, vectorized code in MATLAB for performing analysis
• Data Management and Extraction
• Using high performance computing (parallel programming) to analyze 2 TB of data in short period of time (couple of weeks)
• Writing report and presenting the findings to the VNS group at Boston Scientific

May 2011 – Aug 2011 · 3 mos · Cleveland/Akron, Ohio Area

• At Cleveland Clinic I get trained in:
• Conducting neuroscience experiments on non-human primates
• Building electronic circuits that aid in such experiments
• Using statistical techniques such as ANOVA

Jun 2010 – Aug 2010 · 2 mos · Greater Minneapolis-St. Paul Area

• Worked on understanding the biophysical origins of Local Field Potentials (LFPs) in the Brain. Responsibilities included:
• Gathering information from the literature and colleagues
• Building biophysical based system level, computational models that can produce LFPs with experimentally observed features
• Report and present any new discovery (see patent in the link below) and finding to the group

Aug 2009 – Jul 2015 · 5 yrs 11 mos · Baltimore, Maryland Area

• As a part of Institute of Computational Medicine I worked on developing new techniques for both biophysical based and data driven modeling for studying the brain. The projects that I led are as follows:
• 1. Optimized prediction function for the analysis of unstructured big data
• Used principles from calculus of variation, measure theory, estimation theory, binary quadratic programming, convex optimization, self-concordance and toeplitz matrices.
• Devised the first non-parametric maximum likelihood estimator, the BLML estimator, for estimating densities.
• BLML estimator showed an order of magnitude faster rate for both convergence to true density and computational time than the current gold standards.
• 2. Decoded kinematic features in real-time from neuronal activity in premotor cortex
• Used techniques such as principal component analysis, k-means and hierarchical clustering, maximum likelihood estimation, stochastic optimization, model selection (AIC) and parallel programming.
• Discovered that the spiking activity in premotor cortex is best described by the joint-angles in ellipsoidal regressor-fields at -250ms delay.
• This allowed real-time decoding of hand position and grasp with considerable accuracy (r-square = 0.99).
• 3. Quantified performance limitations of relay neurons
• Used control and probability theory to analyze complex non-linear systems of relay neurons.
• Constructed analytic bounds on relay reliability as a function of the inputs and the system parameters. • This provided the first understanding of the gating mechanisms in relay neurons.
• 4. Studied the effects of deep brain stimulation (DBS) for Parkinson’s disease
• Used distributed computing to simulate a detailed model of network of neurons described by hundreds of non-linear, coupled differential equations.
• Studied the effects of different DBS waveforms on symptoms of Parkinson’s disease.
• Discovered a novel therapeutic DBS waveform which reduced power consumption by 90%.

May 2008 – Jul 2008 · 2 mos · Boston

• Worked in the lab of Dr. Emery Brown for learning statistical techniques in Neuroscience. Specifically, I learned about raster plots, point processes and generalized linear models. I also build point process models to understand the pathology of neural firing in Basal Ganglia in Parkinson's disease.

May 2007 – Jul 2007 · 2 mos · Bengaluru Area, India