John Tanios

John Tanios

About

Detail

Co-Founder/CTO
New York, United States

Timeline


work
Job
school
Education
folder
Project

Rรฉsumรฉ


Jobs verified_user 0% verified
  • Twosense
    Co-Founder/CTO
    Twosense
    Aug 2016 - Current (10 years 2 months)
    Automating identity & multi-factor authentication
  • OneClick Solutions
    Owner
    OneClick Solutions
    Jan 2005 - Current (21 years 9 months)
    Providing custom application development solutions to mid-sized, Fortune 500 companies, and government agencies. Services include web app development, custom LOB (Line-Of-Business) software applications, database administration, networking and technical support.
  • American Red Cross
    Systems Engineer
    American Red Cross
    Oct 2002 - Oct 2004 (2 years 1 month)
Education verified_user 0% verified
  • City University of New YorkCollege of Staten Island
    B.S, Computer Science
    City University of New YorkCollege of Staten Island
    Jan 1999 - Dec 2003 (5 years)
Projects (professional or personal) verified_user 0% verified
  • T
    The Winton Stock Market Challenge
    Jan 2016
    Predicting end of day and intra-day stock returns using previous performance and some masked technical and fundamental indicators.
    Used deep LSTM (Long/Short Term Memory) network to predict returns time-series.
    Achieved WMAE (weighted mean absolute error) of 1728.45. Winning algorithm's WMAE: 1727.54
  • D
    Diabetic Retinopathy Detection
    Jul 2015
    Used deep convolutional networks (CNNs) to detect Diabetic Retinopathy in high resolution retina images.
  • P
    Predicting loan defaults in P2P lending markets
    Aug 2014
    Used random forests to classify and predict borrowers that would default on their loans on LendingClub (peer-to-peer lending platform).
    Achieved AUC of .74
  • P
    Predicting insurance policy purchase for AllState
    Apr 2014
    Used quotes and transaction history to predict whether a customer will purchase an insurance plan, and if so, predict that plan's coverage options.
    Algorithm: ensemble of SVMs + random forests
    Score: 53.282% accuracy. Competition winning score: 53.743%
This is a community-created genome.