SAINAG SHETTY

Software Engineer co-op, Lucidworks Inc.
Raleigh, NC [ Feb 2019 - Present ]

• Part of the Research & Development headed by Dr Chao Han. Working on the Question-Answering system for enriching Solr using Deep Learning technologies
• Rebuilt the entire product module into Docker container. Reduced the environment creation time for deployment
• Developing Python-based API (RESTful Web Service) to expose Deep Learning functionality using Flask
• (Python, Scala, Spark, Docker, Kubeflow, AWS: EC2, Flask)

Researcher, North Carolina State University
Raleigh, NC [ Aug 2018 - Dec 2018 ]

• Researcher under Dr Nagiza Samatova in collaboration with Stanford Medicine
• ◦ Developed an autoencoder model for reconstruction of Brain MRIs and classification
• ◦ Implemented a skull-stripping approach using a u-net architecture to improve the segmentation of tumors
• ◦ Developed a Deep Convoluted Generative Adversarial Network to learn features and perform data augmentation
• (Python, Pandas, Deep Learning, Keras, Tensorflow, CUDA)

Technology Analyst - Developer Intern, ZS Associates
San Francisco, CA [ June 2018 - Aug 2018 ]

• Implemented a semantic parser for Synonym Retrieval for enriching textual data in the insight feed
• Improved data mining tools to provide better business intelligence insights by implementing data science algorithms on Pharmaceutical data
• Collaborated across teams to generate a knowledge base for the development of the domain-specific Chatbot
• Designed and Developed AdHoc query BI visualizations on Tableau
• Used statistical techniques for testing to validate data and interpretations
• (Python, NLTK, AWS, MySQL, Tableau, Pandas)

Software Engineer Intern, Hippily Technology Pvt. Ltd.
Powai, Mumbai [ June 2015 - July 2015 ]

• Analysed the company's system for providing personalised products for the users.
• Developed front-end features and User Interface designs across the Android application of Hippily.
• Added a camera functionality to the application which allows users to take pictures and it is stored on the Hippily server.

Task Scheduling Bot
Guide: Dr. Tim Menzies

• (Node.js, Python, BotKit, Slack APIs, Google Calendar API, MongoDB)
• Developed an algorithm which uses the student's tasks and courses to generate a personalized schedule
• Used Node.js for developing the bot and Python for the scheduling algorithm using MongoDB for the storage

Finding sister cities
Guide: Dr. Micheal Kowolenko

• (NLP, Unstructured Text Analytics, TextBlob, Python, News API)
• Developing an application that takes different parameters such as Demographics, Culture, Economy, Climate and Politics to predict which are the sister cities to the input city
• Using structured data from sources such as Census, and .gov sites. Along with news articles for the unstructured text analytics.

Simulating Game Agent using Q-Network
Guide: Dr. Min Chi

• (Python, TensorFlow, Keras, Q-Learning, Deep Learning, Game Simulation)
• Created an Agent to play the game Flappy Bird with Deep Q Learning, a Reinforcement Learning Technique. 
• Developed a convolutional neural network, trained with a variant of Q-learning, whose input is raw pixels and whose output is a value function estimating future rewards.

Mammogram Classification
Guide: Prof. Jyoti Deshmukh

• Analysed different implementations such as SVM, Convolutional Neural Networks, KNN, Decision Tree, Regression for classification
• Developed an implementation using modified K-means called Adaptive K-means Algorithm.
• MIAS, DDSM datasets were trained and tested against the implementation.

Object Detection and providing purchase links for the object
Guide: Prof. Dnyaneshwar Dhangar

• (Object Detection and Recognition, OpenCV, YOLO, R-CNN, Flipkart API)
• Designed an application that detects everyday objects and took the information to provide purchase links from Flipkart.
• Recognition of the object was done using YOLO framework based on R-CNN. The model was trained on the COCO dataset.
• Provided the user with purchase links for the detected object using the API's provided by the vendor site.

ReadItLater[Webpage Saving System]
Guide: Prof. Preeti Satao, Prof. Priya Parate

• A online news articles and webpage saving system. The article saved is free of all the clutter (like adverts) and is saved in the user's account
• The project used the AlchemyAPI for fetching the article content and store it in the database

Attacks and its prevention on Mail server
Guide: Prof. Savita Lade

• Developed a local Mail Server System using PHP and MySQL.
• Demonstrated various attacks on the System. The various attacks implemented were Buffer Overflow, SQL Injection, Code Injection, Eavesdropping, Phishing and Brute Force Attack.
• Security implementations against the attacks were developed.
• Implementation done using PHP, Javascript, Shell. Ubuntu environment was used and Wireshark was used for Traffic Analysis.

Time Table Creator
Guide: Prof. Venkateshwar Prusty

• An application for generating the timetable for schools/colleges. It takes the required information regarding the course and creates the timetable which can be saved.
• Implementation was done using C++

Attendance

• Android app to manage and record the attendance of a student(college or school).
• The application makes use of the SQLite database to store locally.
• Implementation was done using Android Studio

• Shetty, Sainag, et al. "Real time object detection and recognition and estimation of price of the object.""
  International Journal of Research In Science & Engineering Special Issue 7-ICEMTE March 2017 [PDF]
• Shetty, Sainag, et al. "Detection and Recognition of Objects and Providing Purchase links using APIs."
  International Journal of Engineering Science 10897 (2017).[PDF]