Varun Pathak Email & Phone Number

India

Toronto, Ontario, Canada

• Achieved an 80% reduction in revenue forecasting accuracy by using machine learning and deep learning models that supported the development of long-term financial strategies.
• Built an interactive tool on Streamlit to visualize data drifts, model drift and other insights to communicate the behavior of the model to the business.
• Utilized AWS SageMaker to develop and optimize machine learning workflows, including training jobs and hyperparameter tuning, enhancing the AI team’s model development efficiency and performance.
• Developed continuous ETL workflows to extract, transform, and load data into tables for machine learning models, enhancing predictive analytics capabilities.
• Proactively engaged in strategic planning and task prioritization by collaborating closely with the team members to determine most effective project workflow

Calgary, Alberta, Canada

• Engineered a demand forecasting and mathematical inventory optimization model that potentially optimizes the commercial and retail expenditure by 5%-8% for an agriculture supply cooperative
• Built a machine learning based recommender system that will help increase the operational efficiency of an energy company and potentially save $5 million every year
• Designed pricing dashboard using PowerBI, azure machine learning studio, azure functions and azure synapse to save approximately $8.5 million every day for the retail and pricing team of an Oil & Gas company
• Worked on Natural Language Processing based project to analyze the sentiment towards energy and energy companies that helped the client to make the decision regarding acquisitions
• Programmed multiple end-to-end machine learning pipelines incorporating data processing, data analysis & visualization, feature engineering, model training and validation
• Actively collaborated with the internal teams and business stakeholders in an Agile-Scrum environment

Calgary, Alberta, Canada

• Developed & trained a classification-based deep learning model on image datasets that provides a return on investment of approximately $10 million for an energy company
• Performed exploratory data analysis & ETL on raw datasets to build clean and reusable datasets for analysis
• Communicated technical findings via PowerPoint presentations and Jupyter Notebook to business and technical stakeholders

Edmonton, Alberta, Canada

• Formulated an explainable model for a black-box CNN model with an accuracy of 85%
• Authored reports on explainable AI & rule-based systems under academic supervision to present findings to fellow researchers

Pune, Maharashtra, India

Performed User Acceptance Testing using manual & automated testing methods to provide fully functional client software

Vadodara, Gujarat, India

• Developed a tool to perform Information Retrieval and Data Extraction methods to analyze shortcomings in the codebase and report the findings to essential stakeholders

Master'S Degree, Software Engineering And Intelligent System

Bachelor Of Technology, Computer Science