Skip to content
Saha Rimpa
Saha Rimpa

Saha Rimpa

Director of Data Analytics
Ford Motor Company
Bio

Rimpa Saha is a Director of Data Analytics at Ford Motor Company, where she leads a diverse portfolio of analytics products for the Ford Customer Service Division. Her work focuses on leveraging advanced analytics and machine learning to transform service operations and enhance the vehicle ownership experience.

With over a decade of experience in the automotive and technology sectors, Rimpa has made significant contributions across diverse domains, including Generative AI, predictive modeling, optimization, and statistical forecasting. She plays a pivotal role in architecting scalable, cloud-native solutions that bridge the gap between complex data science and actionable business strategy. As a strategic leader, Rimpa defines the product vision for her organization, empowering and mentoring data scientists to groom young minds in solving real-world business challenges. Her focus remains on building high-performing teams that deliver actionable AI that businesses can seamlessly adopt. Rimpa holds an MS in Financial Engineering from IIT Madras and is a strong advocate for leadership development and diversity in the workplace.

An explorer at heart, Rimpa cherishes spending quality time with her family through global travel. As a toddler mom, she finds great joy in rediscovering the world through the curious lens of a child, exploring new cultures and destinations together.


Hybrid GenAI and ML for Early Quality Detection Using Repair Order Text with an ever Growing Taxonomy

Featuring: Sudharsan SM

This session offers two key takeaways:

(1) a business blueprint for proactively identifying and fixing emerging vehicle quality issues using service data, and

(2) a technical pattern for building a hybrid GenAI + ML system that can classify textual data into an ever-growing set of labels.

You’ll learn how to turn high-volume, messy repair order text into reliable early-warning signals. This approach applies not only to automotive, but to any domain where unstructured text must be continuously organized into a taxonomy that evolves over time.

The technical novelty is a hybrid architecture that combines traditional machine learning with a multi-agent Generative AI workflow designed for an evolving taxonomy, where new failure modes appear continuously. The framework reuses existing categories when the evidence matches known patterns, and creates new categories only when truly novel themes emerge, so reporting and trend continuity remain intact while the label space grows. This makes the session broadly useful for anyone working with textual data who faces a growing multi class classification problem, whether in automotive, warranty, customer support, healthcare, or industrial service.

From a business perspective, the product helps an automotive OEM reduce warranty cost and repeat dealer visits, improve customer satisfaction and NPS, and reduce operational burden across quality, engineering, and field teams by surfacing issues earlier and enabling faster, more proactive fixes.

Professional