KENA SHAH

Experience

• 

  Multi-tech Control Systems

  Feb 2010 - Apr 2013

  Data Engineer (Python)

  Helped automate verification and calibration of automobile and goods-tracking sensors using Python scripts.Analyze performances of implementations of various kinds of sensors using Python scripts. Simulate various test cases, using various Machine Learning techniques like MC simulations, provided by client through FTS (Functional Test Schedule) in Python.Reporting failure of test cases via DR (Defect Report).

• 

  IBM

  Jul 2014 - Jun 2016

  Co-designed an enterprising model for a pharma giant to detect the patients most likely not to adhere to the medication. Provided data-backed proof of concept for counter-intuitive wisdom to the client’s campaign to decipher causes of non-adherence.Employed cutting-edge Forecasting and Machine Learning ensembles to achieve test data True Positive rate to be as high as 84%. Discovering actionable insights for a lending club client pertaining to increasing attrition rate.Provided optimized feature set selecting from 150+ data attributes to help managers detect a potential charge off pro-actively with the test data recall rate of 89%. Used state-of-the-art Machine Learning Techniques like Genetic Algorithms and Random Forest Classifiers for data modelling.

  • Consultant (Data Scientist)

    Dec 2015 - Jun 2016

  • Consultant (Data Scientist)

    Jul 2014 - Dec 2015
• 

  Lancer Insurance Company

  Oct 2016 - now

  Data Scientist

  Mentored 3 Computer Engineering interns in Machine Learning - advised and supervised a Machine Learning pipeline that predicts the occurrence of suit within first 7 days of the claim being reported 80% precision. Designed a pipeline to predict policy-wise liability and physical damage losses and using the output to charge optimal premium for our trucking policies. Designed R Shiny dashboards using data from SQL ETL , external APIs, and Machine Learning backend to provide insight to the underwriters on policies and customers. Working on a Deep Learning intensive project to extract value from the free-form notes taken by the doctors, lawyers and underwriters during the claim management process. Show less