Objective: Building Web application of an Analytical model using NLP, Django, Python, HTML, JavaScript which can identify the similar defects which occurred previously and predict Root Cause Analysis using Machine Learning and suggest the steps to resolve the defects.

Work:

1. Extracting defects that are closed from Jira using API call and building the knowledge repository
2. Extracting new defects from Jira using API call and display them for the user
3. User can click on perform Triage and the suggestions based using NLP are displayed
4. User is given option to approve the suggested triage or for manual override
5. The Approved triage details are updated back to Jira using API
6. Assist with preparation and presentation of dashboards to internal and external demo calls

Impact:

1. Enhanced the time of closing the defects by suggesting the steps to resolve the defects. Defects related to environment reset is automatically done using RPA (which reduced the time for bouncing from 8 hours to 1 hour)
2. Suggestion helped developers to fix the code and database issues
3. Able to solve almost 30% of defects without any manual interventions

Building a Machine Learning model to identify the patterns in the issues using ticket/issue
data for developing bots using Automation tools or to suggest the Knowledge Objects (KO’s).

Preparing training data by removing outliers and performing statistical analysis like correlation index ,Anova ,t-test to extract relevant features.Data analysis and Data Visualization on ticket data by using pandas, matplotlib and seaborn libraries.Predicting the type and count of issues by using regression techniques like Decision Tree Regression.Identifying the root-cause of an issue with the help of LSTM sequence generation techniques.Auto-Trigger the existing bots to resolve the root-cause to provide un-interrupted services

Objective: The objective is to classify the outlined vessel segment as flowing—if blood is moving through the vessel—or stalled if the vessel has no blood flow. Work: 1. Preprocessing the video using OpenCV and forming a bonding box around the region of interest in each frame 2. Converting the ROI to uniform size by Padding using OpenCV 3. Classification of video into flowing or stalled using masked RCNN 4. Increasing the Matthews correlation coefficient using other Neural network models. Impact: 1. Double the analytic throughput of Stall Catchers and could achieve the original goal of analyzing the data 10x faster than lab.