IdentityCognizant, Associate (ML)
Client: Google (DBM) Sep 2018- Present
Objective: To decrease escalation count by predicting whether a ticket is going to meet SLA or not.
â?¢ Built automated tool to navigate tickets to top performing Agent.
â?¢ Joined tables and Cleansed data sets to form consolidated table in PLX SQL database.
â?¢ Clustered Agents based on tenure,submitter location, shift timing, language of ticket
â?¢ Predicted whether the ticket meet SLA or Not and increasing TFS SLA depending upon the dependency.
â?¢ Used K-Means, XG-Boost and Random Forest.
â?¢ Developed and modified business requirement documents in Agile Methodology . Experienced in Google SQL.
â?¢ Packages: Dplyr, Lubridate, Dremel SQL, Random Forest
Client: Apple Aug 2018
Objective: To improve operation metrics of operators by predicting No Action Required cases.
â?¢ Built automated tool to predict whether a ticket is going to be NAR or not.
â?¢ Cleansed and validated data sets based on business requirements.
â?¢ Used Word2Vec and Glove Model to find the relation between the dependent and independent variables.
â?¢ Used Na�¯ve Bayes to perform text classification.
â?¢ Displayed trends and probability of being a ticket to be NAR or not.
â?¢ Used Nltk, Word2Vec, Glove, CNN
Client: Google July 2018
Objective: To find the root cause of increase or decrease in Map Accuracy score and providing insights
to improve operation metrics.
â?¢ Modify databases and tables based on client needs. Developed stored procedures and functions
to modularize the ad hoc queries.
â?¢ Cleansed and validated tables to provide insights in data to increase map accuracy
â?¢ Built dashboards, debugged and explored reasons behind discrepancies in data.
â?¢ Created and Documented test cases to be used for validation of data to satisfy the business
requirements.
â?¢ Packages: Dremel SQL, PLX dashboards
Client: Google Apr 2018-June 2018
Objective: To automate, maintain ,validate and visualize metrics to understand the changes in operation metrics.
â?¢ Automated existing dashboards by creating new temporary tables and stored procedures.
â?¢ Conducted data preparation, and outlier detection using SQL queries and explored the key reasons.
â?¢ Built dashboards with business objective as a primary focus. Conducted data preparation, and outlier
detection using SQL queries and explored the KPI.
â?¢ Packages: Dremel SQL, PLX dashboards
INSOFE, Data Scientist Intern Client: Colombian Power Distribution Unit: Apr 2017- Mar 2018
Objective: Provide Rules to decrease Loss at a geographical location.
â?¢ Clustered meters based on Geographical Location using K Means Clustering..
â?¢ Predicted technical Power loss for clustered meters using Linear Regression and Has provided decision rules to minimize technical loss at a location.
â?¢ Mapped predicted loss to a map-based UI using Folium API.
â?¢ Dumped data into SQL table. Used prediction column as source to display loss on webpage.
â?¢ Packages: : Dplyr, Lubridate,Sqldf, kkmeans ,lm
