OBLIGATION SERVICE 2.0 (HP Inc.) and OBLIGATION EVENT DRIVEN SOLUTION ( HP Inc. via US Services Group)

Worked for Hewelett Packard for almost 18 months. Devloped following things while working for the PSOIT team:

1. Proof of Concept on HP Microservice Acceleration Platform (Powered by AWS).
2. Implemented Redis Cluster, MongoDB as a caching solution by storing whole JSON responses from Source Systems.
3. Developed the async, reactive microservices (Spring Boot) (JAVA Based) performing individual tasks deployable on Netty or Tomcat 9 server.
4. Developed the async, reactive microservices (Flask / FastAPI) (Python Based) performing individual tasks deployable on ASGI or WSGI servers.
5. Dockerized and deployed all the microservice application on HP owned servers managed inside a Kubernetes/Docker Swarm cluster.
6. Developed the logic and deployed an application (Asynchronous) to manage and connect to all the microservices in the cluster.
7. Deployed and managed different tools for functionalities such as Log management, storage and analysis, distributed tracing, api documentation, caching mechanism with Elastic (ELK), Zipkin, Swagger, Redis, respectively.
8. Developed runtime Configuration option and rate limiting for applications without redeployment using Netflix Archaius, PostGre SQL and Bucket4J.
9. Implemented concepts like Bloom Filter to make architecture efficient using Google Guava with Hazelcast and ReBloom (redis Bloom).
10. Deployed and managed HA-Proxy for Redis Cluster, API Gateway and Load Balancing in the cluster.

Mobile Offloading Project

Developed this project for the CSE 535 Mobile Computing Course:

1. Built distributed computing infrastructure for mobile devices using Android Nearby Connections, a P2P networking API. This can be used to do mobile offloading for complex tasks like huge matrices multiplication. Offloading is done here to nearby mobile devices instead of far away machines like fog/cloud servers and thereby saving bandwidth.
2. The infrastructure allows devices to advertise/discover simultaneously. Communication and payload transfer can be done using bluetooth/wifi. This way both pre-connection and post-connection phases are handled using Nearby Connections API.
3. This infrastructure provides low latency, high bandwidth and ability to form fully offline P2P networks. It enables users to switch the type and name of devices at any given point of time. We have used slave devices for computation only until a threshold percentage of batter is consumed.
4. Location, battery level and connection status of slave devices is gathered periodically and failure recovery of slave device is handled in the master device to assign the task to a different slave.
5. Shared Preferences are used to maintain the local state of each server/slave machine at all times.

SAVE THE BLUE PIPT BIRD - DATA VISUALIZATION PROJECT

Developed this project for the CSE 578 Data Visualization Course:

1. Designed an interactive explanatory dashboard that can be used to analyse and visualize important insights from the meteorological and the chemical sensor data to aid Mitch, the student who is investigating the decline of Blue Pipit Bird.
2. Developed this system in response to 2017 VAST Mini Challenge 2 (http://www.vacommunity.org/VAST+Challenge+2017+MC2).
3. The project assisted us in answering three critical questions
   • Characterize the sensors’ performance and operation. Are they all working properly at all times? Can you detect any unexpected behaviors of the sensors through analyzing the readings they capture?Limit your response to no more than 9 images and 1000 words.
   • Now turn your attention to the chemicals themselves. Which chemicals are being detected by the sensor group? What patterns of chemical releases do you see, as being reported in the data? Limit your response to no more than 6 images and 500 words.
   • Which factories are responsible for which chemical releases? Carefully describe how you determined this using all the data you have available. For the factories you identified, describe any observed patterns of operation revealed in the data. Limit your response to no more than 8 images and 1000 words.

Cell Glucose Monitoring (CGM) Prediction Models

Developed this project for the CSE 572 Data Mining Course:

1. This project applies data mining and machine learning techniques on real-time glucose levels obtained from Cell Glucose Monitoring (CGM) sensors of 5 different subjects to identify patterns, recognise if the CGM data is recorded while taking the meal and segment the data based on carbohydrate intake.
2. Tasks implemented:
   • Feature Engineering: Data Preprocessing techniques were applied to the time-series data and four different types of time-series features (statistical, frequency based etc) were extracted by observing patterns in CGM sensor data to form a feature matrix. Standardization and Dimensionality Reduction techniques like Principal Component Analysis (PCA) are applied to the feature matrix.
   • Classification : Trained a model using KNN (K-Nearest Neighbours) and SVM (Support Vector Machines) to classify if a given time-series CGM data indicates the subject is having a meal (1 - Meal) or not having a meal (0 - No Meal) during the respective time frames. Used k-fold cross validation technique to train the model and reported Accuracy, F1 score, Precision and Recall of the machine.
   • Clustering : Trained a unsupervised learning model to cluster the 'Meal Data' (CGM sensor data collected while having a meal) of 5 different subjects based on the amount of carbohydrates in each meal using Means and DBSCAN algorithms. Programming Language : Python
3. Packages Used : scipy, pandas, matplotlib, numpy, sklearn

Female vs Male employment rates in various countries.

Developed this project for the CSE 578 Data Visualization Course - Assignment 2:

1. Designed an interactive explanatory dashboard that can be used to visualize the relationship of male and female employement rates in selected country from 1991 to 2022, in order to observe how this relationship has changed over the time period of our dataset.
2. Lollipop Chart was used for this project.
3. A lollipop chart is similar to a bar chart, but instead of using rectangles to show data values, it shows lines with circles at the top.
4. When the user updates the country value in the drop down, the chart is being redrawn to show the values for the newly selected country.

GDP per Capita in Africa (1960 - 2011)

Developed this project for the CSE 578 Data Visualization Course - Assignment 3:

1. Designed an interactive explanatory dashboard that can be used to visualize GDP per capita in African countries from 1960 till 2011 as per the dataset.
2. The data for this assignment came from the World Bank. It shows, from 1960-2011, the GDP per capita of countries in Africa (based on current USD).
3. When the user hovers or moves the mouse over a country on the map, it higlight the country with a cyan-colored border and show a hover tooltip to the right of the mouse pointer. The tooltip shows the current country and its GDP value.
4. When a user clicks on a country, a line chart is being drawn in this right panel showing the GDP over time for that country from 1960-2011.

Gapminder ScatterPlot

Developed this project for the CSE 578 Data Visualization Course - Assignment 4:

1. Designed an interactive explanatory dashboard that can be used to visualize six different attributes for seven different regions and their countries.
2. In this Assignment, I applied the knowledge of animations and joins to create a scatterplot, that animates as attribute and years for the dataset.
3. The user can change the attributes shown in the scatterplot. I also used animated transitions to update the positions of items.
4. The user can also switch to a different regional set of countries.
5. The functionality of playback was also integrated. A control that lets the user start and stop year playback. When played, the scatterplot begins animating through the years.
6. The current year was also added in background to update in tandem with playback.

TEXT SUMMARIZER USING NLP – MAJOR PROJECT

Developed the text summarizer to fulfil the Bachelor of Engineering degree requirements:

1. Facilitated a web application that handles and condenses a humongous volume of data that is presented in either Text Format, Live Audio Format or Link to a website, by collecting critical information while preserving the original context.
2. Enacted the following tasks Recognition of significant material in source records, identification and elimination of redundant information, key word tokenization, optimization of overview information coverage, and text ordering and rewriting to ensure clarity.
3. In order to summarize lex and gensim algorithm were used.
4. The project flow gows like:
   • Reading from CSV / Reading from Website / Listening from Microphone
   • Data Preprocessing
   • Tokenization
   • Stop Word Removal
   • Deduplication
   • LEX and GENSIM Algorithm

Customer Segmentation using Machine Learning

Developed this project to practice the concepts of Clustering:

1. Implemented a Python-based application that was trained by applying a standard scaler to a dataset and can be used by various department stores to give different incentives and retain customer connections by segregating them based on demographic information and purchasing habits.
2. Executed the following tasks Business Case, Data Preparation, Segmentation with K-means Clustering, Segmentation with Hierarchical Clustering, Comparison of the results from both Clustering methods, Hyperparameter Tuning and Visualization and Interpretation of the Results.
3. This project explores the importance of Customer Segmentation as a core function of CRM as well as the various models for segmenting customers using clustering techniques.

COVID-19 Symptom Tracker Android Application

1. Developed Android application that collects COVID-19 symptoms and stores them on the smartphone sqlite database.
2. End user records a 45 second video by pressing the index finger on the camera with flash on. Using this, heartrate is calculated with image processing techniques to calculate average red intensity using OpenCV Android.
3. After extracting all the frames in the video, difference of average intensity across subsequent frames is used to compute moving average which can be used to deduce average number of peaks in a minute which gives heartrate.
4. Respiratory rate is calculated using accelerometer sensor data by placing on chest for 45 sec.
5. Once sensor values are obtained for X,Y,Z directions for a duration, the duration is broken to lower time window sizes and difference between moving averages of each window is computed to identify the average number of peaks per minute which gives respiratory rate.

Smart Home Gesture Control Application

1. Designed an android application that shows expert gesture tutorials and helps capture/record custom hand gestures. Programmed a Flask server that handles POST requests and stores the user’s custom gesture on the remote server.
2. The developed android application is attached to an ML model for hand gesture recognition.
3. The user can select and practice hand gestures in the app. On saving the final recording into the database the model will return a prediction on the label for the recorded gesture. This gives the user an estimate of how good the user could imitate the selected gesture.
4. Libraries used were TensorFlow, OpenCV, Keras and Numpy with Python 3.6

Exam Remuneration System

Developed this project during my Undergraduate Studies.

1. Developed a gateway through which details regarding the results of any Examination or Re-examination are stored, either for the convenience of the students or the faculty.
2. A major drawback in the traditional way of updating the details is the inconvenience faced by the faculty due to various causes and circumstances.
3. An Evaluation System designed specifically to tackle this helps streamline the process and ease the burden of use. This increases efficiency which effectively increases throughput.
4. The Evaluation System shall include facilities for updating various records such as: Subjects that are valued, Valuation Scheme followed, Details of the Valuator and the final Evaluation Details, to name a few.
5. For the completion of this website, we intend to use technologies like – HTML, CSS, JS, PHP and MySQL.

Technical Forum

Developed this project duriing my Undergraduate Studies.

1. Description: A Web based platform that can be used by customers, employees, and students etc to find the appropriate solution to their problems with the proper guidance from the experts enrolled on the portal.
2. Tools Used: NetBeans IDE, MYSQL Workbench, Apache Tomcat.
3. Technology Used: JAVA, Bootstrap, HTML.
4. Key Learning: Developing a complete project which could be used by many companies for solving the doubts of their customers and allow different users and experts access according to the privileges.

Container Security

Developed this project during my 3rd year of undergraduate studies under the guidance of erudites from HPE (Hewelett Packard Enterprises). HPE-NIE Projectathon : A program started by the Hewelett Packard Enterprises in The National Institute of Engineering, Mysore, India to educate the top 5 percent of the students with the cutting edge technologies.

1. Developed a system for making containers safer by identifying vulnerabilities such as Dirty COW, DROWN Attack, and others.
2. Upgraded the system to assist in the security of the container pipeline, application, container deployment environment, and infrastructure, as well as taking appropriate preventive measures such as integrating with enterprise security tools such as DAGDA, CLAIR, ANCHORE, and meeting or enhancing existing container security policies.
3. Compared the results of various security analysis tools to safeguard the Docker container at varying phases.

Telecom Network Analysis System

Developed this project during my 2nd year of Undergraduate Studies.

1. Description: A System that can be used by customers to find the appropriate plan according to their age group based on their needs.
2. Tools Used: NetBeans IDE, MYSQL Workbench.
3. Key Learning: Developing a complete project which could be used by the Telecom companies for maintaining their data plans and records.

AutoScaling IRAAS over Hybrid Cloud.

1. Designed a web application which uses a hybrid cloud based architecture for recognizing images over a pre trained ML model.
2. Facilitated a web application to condense large volume of data that is delivered in either Text Format, Live Audio format or a URL, by collecting critical information while preserving the original context.
3. Enacted the following tasks: Recognition of important information in source records, identification and removal of superfluous data, key word tokenization, optimization of data, and rewriting to maintain clarity.
4. Deployed the Applications over Openstack (local ubuntu based machine) and Amazon AWS.