-
Experience
Project Engineer | Wipro Technologies PVT. LTD, Pune, India July 2017 - July 2019:
· Experience executing Agile methodologies within a software development lifecycle (SDLC).
· Developed, designed, debugged, unit-tested modules for a customer-facing image processing application.
· Worked in the development of a Report module which involved the rendering of 3D images, 2D images, datasheet, profile control, histogram control etc. using C# and .NET framework.
· Reduced memory usage for application by up to 20% by collecting data from windows performance monitor and optimizing modules consuming high Private bytes memory.
· Led and mentored a team of new graduates, creating modules to track their work performance and conducted a periodic review of coding standards.
· Devised training on software architecture patterns and concepts of object-oriented programming to 10+ new hires.
-
Projects
· SIMPLIFIED AMAZON DYNAMO USING ANDROID BASED SYSTEM CONTAINING 5 ANDROID VIRTUAL DEVICES (AVDs): Focus of this project was to provide both availability and linearizability simultaneously. The network was partitioned into 5 nodes corresponding to 5 AVDs and Chain Replication was implemented to store replicas of data in the networking nodes. Replicas were used for failure handling and recovery such that normal working is not interrupted. The project was implemented using Java in Android Studio.
· SEARCH ENGINE FOR TWEETS (Link to demo: https://youtu.be/fcpt8srpb_M): Developed a python script to crawl Twitter for tweets of influential people based on various criteria such as language, location etc. Processed JSON files and indexed in Solr for querying, tuned the performance of different IR Models and evaluated its performance using trec_eval primarily wrt MAP values and created an interactive UI summarization of different topics and hosted the server on AWS. Users could query and see the relevant tweets and insights derived from those tweets e.g. sentiment analysis. Built with Python, Tweepy, Twarc, Solr, NLTK, SK Learn, Django, JavaScript, HTML, Angular JS, AWS, EC2.
· IMPLEMENTATION OF MACHINE LEARNING ALGORITHMS USING PREDICTIVE ANALYSIS: Implemented Supervised learning algorithms KNN, Random Forest, Dimensionality reduction using PCA, Unsupervised algorithm K-Means from scratch using numpy and pandas. Supervised learning algorithms SVM, Logistic Regression, Decision Tree, KNN using Scikit-learn library and Ensemble model using the voting classifier of the above-mentioned algorithms using Scikit-learn library in Python.
· PREDICTIVE ANALYTICS WITH APACHE SPARK FOR BIG DATA: The objective of the project was to implement a multi-label movie genre classification model using Apache Spark. The model was trained on different features extracted from the plots of movies. 3 different feature extraction methods were used for feature engineering: term-document matrix or term frequency, TF-IDF and Word2Vec. These features were used to train a model that was in turn used for predictions. The models used were Logistic Regression and Support Vector Machine (Linear SVM). Built with PySpark in python.
· IMPLEMENTATION OF MAP-REDUCE BASED CODES FOR BIG DATA ALGORITHMS IN HADOOP: Implemented Inverted index, counting the Top 10 occurring trigrams in the given corpora, joining 2 tables using a primary key and KNN to predict the labels of a given test set using Hadoop-streaming in Python.
· DISTRIBUTED CHAT APPLICATION FOR ANDROID: Developed an android chat app that can multicast simultaneous messages and maintains Total-FIFO ordering in 5 different AVDs using socket programming in Java. Messages were sent over TCP network protocol.