nidhi davawala

Shaan

           

About Me

I am a Computer Science graduate at University of Massachusetts Amherst. Currently I am interning at Goldman Sachs on their alert management system. Through my courses and projects, I found interest in the area of Machine Learning and love its multidiscplinary nature. I have previously interned at the Indian Space Research Organization, wherein I explored Computer Vision in radar imagery. At leisure, I love to read!

Internships

Goldman Sachs, New York Ongoing
  • Working on the Enterprise Health Restoration Services for alert management
  • Generating interpretable user manuals from JSON worklows and integrating it across the system heath check platform.

Space Applications Center, Indian Space Research Organization Summer'18
  • Conducted supervised image classification on agricultural land using RADARSAT-2 Synthetic Aperture Radar(SAR) Data
  • Implemented Wishart Classification for classifying an image into varied crops, water bodies and populated regions
  • Generated change map using hypothesis testing to identify the areas with significant changes over a period of time
  • Conducted analysis of changing trends in crop lifecycles; useful to farmers for crop-yield prediction

Publications

N. Varia, N. Davawala, S. Chirakkal, D. Haldar, R.Ghosh, D. Putrevu
ISPRS 2018

Projects

  • Implemented a mini-search engine capable of handling HTTP query requests to retrieve webpages from keywords.
  • Used Hadoop File Systems to store the files and corresponding URLs; Apache Spark for generating inverted index.
  • Answered queries by retrieving the corresponding files stored as key-value pairs on RocksDB.
  • Link to Github repository
  • Developed an interactive website using Boostrap for multi-view D3 visualizations of OSMI Mental Health in Tech surveys
  • Performed data analysis on surveys for 3 years to understand the changing trends of mental health of employees in IT industry.
  • Gained insights on the cause and spread of mental disorders, evaluated the support system and made suggestions to improve it.
  • Link to demo
  • Used deep learning models to predict the presence and sub-type of hemorrhage from medical cranial images
  • Trained a Convolutional Neural Networks model (CNN) with a prediction accuracy of 91.3%.
  • Improved model classifcation accuracy using Recurrent Neural Networks layered with CNN.
  • Developed a mechanism for automatic background subtraction to identify the moving objects in a video
  • Generated a basis vector of background using Principal Component Analysis and Locality Preserving Projection
  • Subtracted the modelled background from the test video to detect the foreground with 97.7% accuracy.

Skills

  • Programming: C/C++, Python, HTML
  • Machine Learning: Numpy, Sklearn, TensorFlow
  • Tools: Hadoop, PostgreSQL, LaTeX
  • Neural Networks (NTU Singapore)
  • AI-Search Methods for Problem Solving(IIT Madras)
  • Chairperson,Women in Engineering (WIE) Affinity Group of IEEE Student Branch, DA-IICT (2018)
  • Executive Committee Member of IEEE Industry Applications Society DA-IICT Student Branch Chapter (2016)