SWAPNALI GUJAR

As an accomplished Principal Data Scientist, I am motivated by challenges and driven by data, delivering meaningful AI/ML/Gen AI based solutions to the customers; leading AI/ML, data science, data analytics-based projects from Proof of Concept to Production, by closely working with the business, to drive success.

Principal Data Scientist, Cummins Inc, 06/2017, Remote, USA

Cummins Engine Prognostics Modeling: With expertise in leading the development of a robust data science and feature engineering pipeline, I have successfully enabled the delivery of advanced statistical and machine learning models. My instrumental role in preventing catastrophic failures of Cummins Diesel Engines in the mining industry resulted in significant cost avoidance of $20M for Cummins and valued customers. Leveraging predictive analytics algorithms such as multi-linear regression, KNN, random forest, decision tree, and XGBoost, I have effectively analyzed time series data to accurately forecast potential engine failures. 

Relevant Service Request Recommender: I have architected a data science and data engineering pipeline to deliver an end-to-end recommendation engine for Cummins Field Service engineers. By utilizing textual data, this innovative solution has reduced the closing of service requests by an average of 11 days, generating a value of $0.5M within just 6 months. As part of this project, I have shouldered various responsibilities including recommendation model development, multi-class classification model development using NLP, Gen-AI, and machine learning techniques, business communication, project entitlement, value derivation, and coaching and mentoring junior data scientists. 

Cummins Product Reliability Analytics: My expertise extends to leading reliability data management and reliability prediction for Cummins products. By reporting to regulatory organizations and enabling business units to perform financial planning for warranty purposes, I have demonstrated proficiency in core statistical product survival analytics methods such as Weibull. 

HR Data Analytics: My contributions include developing HR analytics-based models to predict attrition rates and conducting descriptive analytics using exit interview data from ex-employees. 

Other Professional Experience:

Project Lead at various organizations such as Harman Internationa, Tata Consultancy Services, Cognizant and KPIT (July'2004 - June-2017)

Led multiple software engineering projects that include development of software & controls, test automation framework in engine, powertrain and connected car domains.

• Microsoft Azure Foundations, #H867-6488
• Purdue Badge Program Certification in Data Science II
• Neural Network and Deep Learning, 01/09/21, #Z33UF79M2ZLV, deeplearning.ai
• Machine Learning Using Python Certification, 03/15/20, #6KZMS9K8QP5M, IBM
• Architecting software for Smart Internet of Things, 09/01/19, #J9UQ78E4FD4N, 94%
• Industrial IoT Markets and Security Certification, 03/01/19, University of Colorado Boulder, 97%

• Led the successful delivery of AI/ML end-to-end solutions that generated over $25M in value.
• Primary inventor of a patent on a data science-driven project titled 'Systems and Methods for Determining Exhibited Useful Life of Sensors in Monitored Systems (NP US),' resulting in Patent Number: 11,959,433.
• Received first prize in Indiana University's data-thon and hackathon 2023 on hate speech, developing NLP models for predicting racial bias and identifying bias in social media content.
• Filed multiple Intellectual Property Inventions and Trade Secrets based on data science projects.
• Received the 'Customer Success Award' and the 'Data Science Innovation Award' at Cummins.
• Served as a champion of employee volunteering in community engagement efforts.

Academic Projects (M.S. In Data Science):

• Home loan Credit Defaulter Prediction using Machine learning & Deep Learning
• Jeopardy Game Training Game Development using NLP and Recommender Model
• "WhatsMyWorth": Data Management & Visualization for "AI & Data Field Related Salary Insights"
• Visualizing Human Genetics Disorders Using Parents' history