Piyush Patil

• Refactored and modularized legacy ML codebase into a scalable, clean architecture for predicting battery cranking ability. • Implemented a cranking curve preprocessing module with first order signal gradients, boosting prediction accuracy by 10%. • Utilizing Azure DevOps for version control, task tracking, and CI/CD pipeline integration. • Testing and validating the new pipeline on real-world battery data in Snowflake, enabling scalable model evaluation.

• Developed an intent recognition system with Azure OpenAI and custom taxonomies to classify user queries in conversational AI. • Utilized LLMs and Generative AI for intent-based query reformulation, improving response relevance. • Focused on enhancing user engagement conversational AI responses through intent-driven insights.

• Analyze driving dynamics of test vehicles using G-Force trackers to understand their influence on motion sickness development. • Develop automated evaluation systems for GPS measurement data to streamline analysis processes. • Utilize machine learning techniques to construct predictive models for motion sickness based on g-force data. • Collaborate with interdisciplinary teams to integrate findings into innovative automotive technologies.

• Developed a machine learning model to classify structural static load conditions, enhancing predictive analysis. • Processed datasets from multiple sensors and applied various algorithms for optimal performance. • Visualized model performance and results effectively using Matplotlib. • Achieved 95% accuracy and robustness in detecting 24 diverse load conditions, enhancing data-driven insights.

• Managed and interacted with databases to retrieve, store, and update data on Oracle. • Utilized Oracle’s Siebel tools to create, modify and manipulate data objects. • Created and Maintained documentation describing data objects, their relationships, and their use in analytics and reporting. • Designed and Configured Siebel objects, including Screens and Appletes.

• Conducted data analysis and documentation to track manufacturing performance, helping to identify trends and areas for improvement. • Collaborated on the design and development of tread manufacturing processes for truck and tractor tires, contributing to the creation of efficient and high-quality tire treads.

Relevant Coursework: Machine Learning, Unsupervised Deep Learning, Introduction to C and C++, Embedded Systems, Mechatronics System Design, Project Management

Relevant Coursework: Mechatronics, C, C++