Ujjwal Peshin

I am a Masters of Data Science student at Data Science Institute at Columbia University , graduating in December 2019. I am looking for exciting internship opportunities where I can apply my Machine Learning and Data Science skills for the greater good. I have a strong Computer Science and Data Science intersection background.

I have worked at ZS Associates as a Data Science Associate for a year, where I majorly worked on Constrained Optimization, State Space Search and Imbalanced Classification problems in healthcare domain using Read World Evidence data. I graduated in Computer Engineering from Netaji Subhas Institute of Technology under the University of Delhi, which is now Netaji Subhas University of Technology, in 2017. I worked with Prof. Sushama Nagpal for my Bachelors project, Touchless typing for Charcater Recognition.

Email  /  Resume  /  LinkedIn  /  GitHub

Work Experience

Data Scientist, ADS and RWE Team ( ZS Associates )
July '17 - July '18

My role was to develop and use advanced data science techniques to solve difficult client problems in an efficient manner. Some of the projects are as follows:

  • Detecting patients at high risk of developing AFib in the next 3-6 months Produced multiple classification models to find prevalence of AFib using RNNs and advanced ensembling techniques, providing a potential upscale of $30 million to the client.
  • Clinical Trial Patient Pool Optimization Formed a state space search model to optimize Inclusion/Exclusion patient pool for clinical trials, which was included in a pipeline to develop a complete clinical trial solution.
  • Physician Record Visualizer Prepared a deliverable using Django to display physician records using various data sources to provide an enhanced view to the client regarding each physician record.
  • Detecting ADRs using CNNs Developed a model to detect ADRs from tweets using CNN.
  • Teaching Assistantships
    teaching

    Teaching Assistant: Elements of Data Science, COMSW4995 - Fall 2018 [Columbia University] with Prof. Bryan Gibson





    Homepage Credits: Thanks a lot, Jon!