My educational and professional background

I have a Ph.D. in EE from the Department of Electrical and Computer Engineering at the University of Illinois at Chicago and a B.Tech in Instrumentation Engineering from the Indian Institute of Technology (IIT), Kharagpur.

Currently, I work as a Senior Principal Engineer at ON Semiconductor in the Low-voltage Device Design group.

Here is my LinkedIn profile.

I was elevated to the grade of Senior Member of IEEE for my contributions towards power electronics in 2015. I have authored/co-authored more than 25 peer-reviewed Transaction and Conference papers, 2 monographs/book chapters, and 4 U.S. Patents. I also serve on the technical program committee as Track/Topic chair in many IEEE conferences.

My experience with data analytics/machine learning

My transition from a 'user' of tools to a hands-on data science practitioner

Throughout my career, I have worked with complex, high-volume, and high-dimensional scientific and engineering data. I have extracted and analyzed large datasets, created advanced visualizations, and presented insights to my technical teams. But in all those endeavors, I was largely dependent on a fixed set of tools and was never exposed to the true power of underlying mechanics of data analysis.

However, since last 2 years, I have become keenly interested in the tools and algorithms which work under the hood and this has opened up a world of possibilities for me. I have discovered the joy of not depending on enterprise tools and writing my own functions and codes in languages such as Python or R. This has given me a much broader perspective and much sharper insights into any data related problem. I have actively started contributing to open-source communities by sharing my analytics projects on Github and social forums. In parallel, I have started learning about core algorithms and how they solve the range of analysis and machine learning problems. I am appreciating the whole process deeply and feeling empowered to peel back the layers. Today, I am able to think about a data problem from multiple viewpoints and analyze/model/learm from it with high degree scientific rigor using custom code and algorithms.

Continuing education in data analytics and machine learning

I have audited multitude of courses on statistics, algorithms, machine learning, and analytics on online platforms such as Coursera and edX. Some of the prominent ones are listed here

  • “Artificial Intelligence: An Introduction To Neural Networks And Deep Learning”, Stanford Continuing Studies, Stanford University (on-campus)
  • “Data Science: Data to Insights”, Massachusetts Institute of Technology, MIT Professional Program
  • Machine Learning (Coursera - Stanford University)
  • Data Science MicroMasters program (edX - UC San Diego)
  • Deep Learning specialization (Coursera - deeplearning.ai)
  • Machine Learning with Case Studies Approach (Coursera – University of Washington)
  • Algorithm Design and Analysis (edX – Penn State University)
  • Artificial Intelligence (edX – Columbia University)
  • Bayesian statistics - Techniques and Models (Coursera – UC Santa Cruz)

    • Apart from these, I am registered to start a Masters of Science (MS)Degree in Analytics (specializing in Computational Data Analytics) from Georgia Tech this Fall (August 2018).

    Applying analytics and machine learning to my domain

    I have been working on following projects, related to machine learning, in my domain,

  • Python-based data analytics framework for semiconductor manufacturing: Using Python-based Jupyter notebooks for analyzing silicon wafer processing data. This involves failure bin analysis, statistical plotting, normality analysis, quantile plots, creating parametric maps out of positional data, bi-variate regression model builds, causality discovery by multi-variate regression pipeline, outlier detection with Gaussian mixture models, etc.

  • Semiconductor device design automation pipeline with machine learning: This project aims to automate/aid the complex technology development and device design tasks in the field of high-power semiconductors, using machine learning and statistical modeling. Goal is to develop an integrated software platform which can control various Technology Computer-Aided Design (TCAD) simulation programs, extract the relevant outputs from those simulations, and build a design guideline for optimal performance and reliability in a given power management application.

  • Deep learning-based semiconductor design feature extraction: Goal of this project is to use deep learning framework to mimic ‘high-level’ design experience of domain experts (in the field of semiconductor processing, device, and packaging) by classification of designs into categories such as ‘sub-optimal’ or ‘aggressive’.

    • Consultancy with a AI-based startup

    Follwoing my passion about AI/ML/data science, in my spare time, I work with a AI-driven starup in Palo Alto, CA as a Consulting Scientist. My responsibility is to advise on generative models for software test scripts and test data, and to work with the Natural Language Processing (NLP) team to discover hidden patterns in the test scripts of various web-based applications with the eventual goal of automating the testing process to the largest extent possible. I intend to apply statistical modeling techniques (such as Gaussian mixture-models, frequency counters, topic modeling, variational autoencoder, etc.) along with text-mining algorithms to generate synthetic test scripts and then process them through a deep learning model to obtain highly accurate automatic test case synthesis from minimal seed data set.