Mukta Das

AI Strategy, Agentic AI, and GenAI Initiatives.

• Led the design of Tata AI Service Advisor, an agentic AI-powered assistant to transform automotive service operations from reactive diagnostics to predictive maintenance, reducing turnaround time, and improving service quality.
• Defined an end-to-end AI strategy and architecture, addressing the challenges of fragmented technical knowledge, inconsistent technician expertise, and increasing vehicle complexity.
• Designed a multi-agent architecture (Diagnosis, Repair, Spare Parts Agents) enabling autonomous collaboration, reasoning, and decision-making across service workflows.

Adaptive step-by-step multimedia repair guidance.

• Integrated Repair Agent with Diagnosis and Spare Search Agents via orchestration layer for seamless decision-making.
• Implemented self-learning feedback loops, enabling continuous improvement through historical repair data and outcomes.

CKD Platform – Digital Twin and Computer Vision.

• Designed a next-gen CKD (Complete Knock Down) digital platform to streamline kitting and packing processes using AI and computer vision.
• Proposed digital twin-based inspection system integrated with CAD–PLM data for real-time validation.
• Defined n-tier scalable architecture using open-source frameworks for enterprise deployment.
• Enabled reduction in manual errors, improved inspection accuracy, and operational efficiency.

Enterprise AI Transformation and Agentic Frameworks.

• Contributed to enterprise-wide agentic AI strategy, focusing on scalable, modular, and self-governing AI systems.
• Applied principles of autonomous decision-making, multi-agent collaboration, and continuous learning systems.
• Designed solutions across dealer operations, diagnostics, spare parts, and repair ecosystems.

AI Thought Leadership, Whitepaper Contribution

• Developed an agentic AI transformation blueprint inspired by BMW's AI strategy, outlining enterprise adoption via a hybrid CoE model.
• Defined roadmap for scalable, ethical, and competitive AI deployment in automotive ecosystems.

• Developed demand forecasting model for a petroleum industry for various products, across different time horizons and granularities such as geographical area, population cluster, supply points using R programming Studio.
• Developed demand forecasting model with deployment of COVID intermittent pattern recognition, for various products of US based FMCG client to optimize the inventory at warehouse level using SAS Viya and forecasting techniques like ARIMA, SARIMA & time series techniques.
• Analyzed complex data and conducted a price variance analysis to find the root causes of inefficiencies, bottlenecks, or deviations in the supply chain processes for the US based FMCG Client.

• Recommendation Engine - Offer recommendation to suggest personalized offers to customers based on their usage patterns, demographics, and other relevant factors by Collaborative filtering machine learning technique using SAS Studio packages and R programming Studio for model building.
• Evaluated suppliers to maintain cost controls and improve operations by applying cross-sell propensity models using Pandas, Keras, matplotlib.
• Developed multiple PoCs in order to check the existing sales pattern and eventually deploy analytical platform for the specific scenarios.
• Analyzed complex datasets in order to derive meaningful business insights and integrate the same in a visualization tool - Tableau & Power BI.

• Developed predictive models to analyze customer behavior and improve product recommendations.
• Utilized machine learning algorithms such as random forest, logistic regression, neural networks, and support vector machines.

• Developed a statistical model using Logistic regression to standardize the service marketing strategies through proper targeting of chassis for scheduled service and pitch to the customers with chassis who is more likely to come for a scheduled service in next month rather than random targeting & assign probability score using SAS Eminer & Base SAS.
• Developed a statistical model to identify chassis which are likely to churn out of service - meaning not generating service revenue anymore so that action can be taken on those chassis to stop the churn using SAS Eminer & Base SAS.
• Developed a lead scoring propensity model using Random Forest to predict the lead getting converted to retail.
• Conducted a micro market analysis to identify the high probable micro sales area for Passenger Vehicle customers and its integration with Tableau.
• Developed a forecasting & causality model for predicting Total Industry Volume for Commercial Vehicles at regional & PAN India level using SAS Forecast Studio with techniques like ATIMA, Univariate time series.