Anand Ramakrishnan

Experience

• 

  Soft Robotics Lab WPI

  Aug 2015 - May 2017

  Research Student

  • Performed Deep Learning for a Soft Force Sensor Array (Thesis project)• Created a soft actuator-sensor matrix to study palpations of an organ • Creating Soft Sensor to measure the curvature of Bending objects in 3D• Creating Soft Sensor to measure force applied both in shear and Normal directions

• 

  Worcester Polytechnic Institute

  Aug 2017 - now

  • Developing Architectures that learn to predict the quality of teacher-student interactions by parsing videos using CNN LSTMs.•Working on models that performs Face Detection using Mask RCNN and FRCNN•Building architectures that perform Multi Object Tracking using LSTM attention models•Creating multi task learning networks that identifies emotions and age.•Worked on analyzing how much error is involved when a detector is trained to approximate an existing measurement•Involves working with Amazon S3, Rekognition and sending out hits onto MTurk. •Architectures being built on Pytorch, Caffe and Keras. Show less

  • Research Assistant

    May 2017 - now

  • Teaching Assistant

    Aug 2017 - Dec 2018
• 

  Bosch Center for Artificial Intelligence

  May 2018 - Aug 2018

  Deep Learning Intern

  . Implemented facial keypoint detection using stacked hourglass architecture on CelebA and AFLW datasetsachieving Normalized Mean Error of 1.62% and 4.20 ± 0.41% respectively.. Created a novel Heatmap based uncertainty metric to provide error bounds.. Evaluated metric using active learning literature.

• 

  PARC

  May 2019 - Aug 2019

  Research Intern
• 

  Norfolk Southern

  Jan 2021 - now

  Deep Learning Computer Vision Scientist