Rojin Varghese

AI Platform & Architecture Leadership: Designed and implemented a robust Data Science and AI platform utilizing Azure
services, establishing the foundational architecture for the companys Data Science, Machine Learning, and Generative AI
initiatives. Served as the AI hub architect, developing comprehensive solution architectures and ensuring strategic alignment
with key stakeholders across the organization.
Alternate Parts Identification (RAG, OpenAI & Agentic Framework): Led the deployment of an innovative AI tool
leveraging OpenAI and Agentic frameworks to identify alternatives for single-sourced or obsolete parts. Automated parametric
data extraction and report generation, streamlining alternate selection and achieving approximately $2.5 Million in annual
savings. Developed a scalable solution critical for new product development and proactive supply chain risk management.

Manufacturing Operations Troubleshooting Chatbot (RAG Framework): Developed and deployed a critical
troubleshooting chatbot for manufacturing production floor technicians and assemblers. Leveraged a RAG framework,
indexing historical troubleshooting data, manufacturing escalation records, and field escalation records to provide accurate
and relevant root cause suggestions. Implemented a robust system to recommend troubleshooting steps when a confident
root cause cannot be provided, ensuring continuous support for issue resolution. and Recommendation Tool (Gen AI): Designed and implemented a Generative AI
powered tool to retrieve relevant boilerplate designs in response to user queries, providing engineers with valuable
references to accelerate project and reduce iteration time.

Team & Capability Development: Led, mentored, and upskilled a high-performing team of 12 Data Scientists, fostering
their professional growth and technical capabilities.
Data Strategy & Integration: Orchestrated the synchronization of multiple complex data sources and pipelines to establish
a unified platform for corporate analytics, enhancing data accessibility and reliability.
AI Spearheaded the design and architecture of advanced AI solutions, evaluating over 7 predictive
models to uncover critical new business insights and drive data-driven strategies. Provided key stakeholders with consistent and actionable data insights, facilitating
informed decision-making and ensuring the timely delivery of critical analytical support.

Clearance Forecasting and Markdown optimization Engine
• Designed data pipelines for the models for different measurement tests and deployed base demand prediction models,
Markdown optimizer models, and elasticity calculation models.
• Optimized models using time series techniques and performed statistical hypothesis tests on data for validation and
measurement.
• Planned and executed multiple A/B tests to measure the performance of different models.

Machine Learning and Data Science Integration: Applied machine learning algorithms and data science techniques to
optimize processes and derive actionable insights from collected data.
Operational Analytics: Leveraged operational analytics to streamline assembly and distribution processes, reducing
workforce requirements and associated costs.

Customer Email and Chat Analysis: Utilized Natural Language Processing and various machine learning techniques to
analyze customer emails and chat details, extracting valuable insights.
Social media/Web/Text Analytics: Harvested data from social media platforms like Facebook and Twitter using web
crawling techniques. Analyzed this data with advanced data mining techniques to produce actionable insights.
Development of Analytics Tools: Created tools to monitor online website metrics, providing insights such as customer
flow, time spent on pages, and sales versus visits. Employed machine learning algorithms to enhance these analytics.

Email Classification and Categorization using NLP: Designed and automated a machine learning model to classify an
insurance clients emails into 20 distinct categories, ensuring accurate forwarding to respective departments. Achieved an
impressive 85% accuracy rate.
Analyzing Unstructured Big Data: Collaborated on a groundbreaking concept to analyze Big Data directly without prior
structuring, significantly reducing processing time. This innovative work was published in the international conference
ESOMAR 2013.

Played a key role in developing both static and dynamic trip
planners, as well as a GPS-based real-time trip planner and Route Information System, leveraging advanced Machine Learning
techniques.