SUDARSHNA PATEL

Masters Thesis: A Reinforcement Learning based optimization for a real-world application Drinking Water System (Aquasuite OPIR). Developing a RL model to control the water production which can help optimize the water flow inside reservoirs for a demand forecast based on certain hydraulic parameters.

Architected, designed and developed​ backend for Family Doctor & Coupon System and led implementation of RESTful APIs for the same. Integrated real-time chat service using Ably between customers and doctors for doctor consultations. Optimized APIs for faster response time using caching. Designed and implemented database architecture for various projects in Postgres. Experienced in a​pplication quality assurance, testing APIs and f​unctionality utilizing TDD principles.

Developed web crawlers for various websites using Beautifulsoup, Selenium, Scrapy and Python.

Development of Image upload Website: P​ roject involved development of a website to collect images from users using php and mysql in laravel with html and javascript framework and hosted it on university server.

Eczema Detection Desktop App:​ ​Project involved development of GUI for an automated eczema detection system using C++ and OpenCv on Qt platform.

Courses Taken: Machine Learning: Supervised/Unsupervised Learning, Transfer & Ensemble Learning, Time series analysis Deep Learning: DNN, CNN, RNN, Transformers, Autoencoders, GAN Data Mining: Data Visualization (PCA), anomaly detection, Recommender Systems (Matrix Factorization) Text Mining: Text categorization, text as a sequence, sentiment analysis, RNNs, Transformers, BERT) Social Networks Analysis Reinforcement Learning Game AI Information Retrieval Evolutionary Algorithms

Master's thesis: Sign Language Recognition Using Motion Information - Developed a Sign Language Recognition System that was able to recognize dynamic gestures in a video using Motion Gradient Orientation images and SVM classification algorithm and the final model obtained an accuracy of 66.25% on ASLLVD dataset.