Rishabh Sahrawat

> Mainly working in the field of Natural Language Processing > Built, deployed and continuously maintaining Search Suggestion Service using Convolutional Neural Network, to provide suggestions (companies) for user provided keywords > Built data preprocessing pipeline for large Crunchbase database > API for handling multiple user requests simultaneously using Flask & Redis > Customers include Daimler AG, Miele etc. > Written Android Application for deploying Search Suggestion model on mobile devices

Research on different Generative Adversarial Networks

Approach to Combine BEGAN and CycleGAN >Study literature on image-to-image translation >Implemented CycleGAN for translating images from one domain to another >Applied CycleGAN on many different created and already available datasets >Improved translation results through adding a new loss function >Combined image-to-image translation approach with synthetic image generation model called Boundary Equilibrium GAN >Learned to use a deep learning python-based framework PyTorch

Deep Learning in Image Processing >Learned deep learning theory on Generative Adversarial Networks (GANs) >Learned to use deep learning python-based framework Theano/Lasagne >Implemented several GAN based approaches for generating synthetic images

Assistant at Chair of Digital Libraries and Web Information Systems >Study literature on Reinforcement Learning >Use of Reinforcement Learning technique with Deep Neural Networks >Implemented network based on reinforcement learning to train an agent for playing a text-based game >Learned to use python based framework Keras/ Theano

Machine Learning and Context Recognition >3 Projects using Machine Learning Algorithms >Indoor Localization using WiFi Fingerprints- Implemented various supervised classification algorithms >Spotting Drinking Gestures from Wrist Motion- Learned Pattern Recognition concepts using Hidden Markov Model >Motif Discovery- Learned Motif Pattern concepts using Segregate Approximation Algorithm

Text Mining Project- Sentiment Analysis on Twitter data based on BREXIT >Carried out various pre-processing techniques on the data >Implemented different supervised classification based algorithms like Support Vector Machine, Naive Bayes, Decision Tree etc for classification >Implemented Cross Validation techniques for better modal predictions

Field of Web of Things >Understood working of Festo MPS Workstation >Replaced the I/O Simulator of the workstation with NodeMCU >Learned the functionality of NodeMCU >Operated the workstation through a mobile device via NodeMCU

Intelligent Systems; Machine Learning and Context Recognition; Text Mining; Artificial Intelligence; Deep Learning