Vigneshwar M Email & Phone Number

Bengaluru, Karnataka, India

Experienced in applying machine learning, statistical pattern recognition and Data mining algorithms with accuracies of more than 85%.Performed Natural Language Processing (NLP) using algorithms like K-means and LDA modeling in text data of incident tickets to derive clustering groups and used TSNE to project them as data points to provide insights that led to a 50% increase in ticket closure.Used bidirectional LSTM as a classification algorithm for predicting the class ID of parts with different features.Used Dataiku DSS to build a complete visual flow for end-to-end processing of data from preprocessing to scoring the model with more than a 0.9 AOC score.Built a visual flow in Dataiku DSS for implementing rules with predictions using XG Boost.Performed statistical analysis on classification models to evaluate model performance using various metrics to understand the feature importance.Performed data cleaning and feature engineering for various projects having big data using spark in Databricks.

Bengaluru, Karnataka, India

Established connections to Snowflake from Databrics, Dataiku and Jupyter Notebooks to query warehouse data in the database using advanced SQL and Python using Snowpark.Preprocessed the data using Python and performed Exploratory Data analysis to create visualizations with Pandas, Seaborn, and Matplotlib and derive insights. Established connection with Snowflake tables from Power BI for visualizing insights.Performed Bi, Tri -gram analysis for the clean text data after pre-processing to provide insights on repeated incident tickets and derive trends of data over some time.Developed interactive web application using Streamlit library for presenting to stakeholders.Developed code using Python in Azure Data bricks for ETL processing of files in different formats.Scheduled and monitored the jobs in Data Brics and implemented bug fixes whenever necessary.Migrated Datastage ETL pipelines to Snaplogic and performed unit testing.Developed a peer-review automation tool customized to detect errors in SQL queries that reduced manual effort by almost 75%.