AI and Model-Based Design in EV Engineering: An Interview with Vijayalayan R | MathWorks

In an interview with EVreporter, Vijayalayan R, who leads Automotive Industry and Location Field Application Engineering at MathWorks India, discusses how AI and Model-Based Design are influencing automotive engineering. He shares insights into their applications in EV battery management, software-defined vehicles, autonomous driving, and engineering workflows, highlighting the role of simulation, testing, and validation in developing reliable automotive systems.

How is the combination of Model-Based Design and generative AI changing the speed and rigor of engineering design, from concept through validation?

Generative AI can accelerate engineering workflows, but simulation, verification, and traceability remain essential to delivering high-confidence designs.

Model-Based Design enables engineers to use a system model throughout the development lifecycle, from concept exploration and simulation to implementation, testing, and validation. Generative AI capabilities such as MATLAB® Copilot and Simulink® Copilot help engineers work more efficiently within that workflow by assisting with activities such as model development, coding, debugging, documentation, and design exploration.

Beyond AI assistance within engineering tools, many organizations are also exploring workflows that allow AI agents to participate in engineering tasks across the development lifecycle. Using technologies such as the MATLAB Agentic Toolkit and Simulink Agentic Toolkit, AI agents can interact with engineering models, simulations, tests, and requirements workflows within their trusted existing Model-Based Design processes. This enables teams to automate repetitive engineering activities, accelerate design exploration, execute validation workflows, and improve productivity. Engineers continue to review results, make decisions, and maintain accountability for system performance, safety, and compliance. The objective is to create AI-augmented engineering that combines the speed of AI with the rigor of simulation, verification, validation, and traceability.

How does mixing human ideas with AI insights redefine the role of modern designers, and how does AI bridge the gap between theory and a practical, physical solution?

AI can help engineers explore ideas faster, but engineering judgment remains central to turning those ideas into real products.

As AI becomes part of the engineering workflow, the role of engineers increasingly centers on system-level thinking, trade-off analysis, and validation. AI can speed up the exploration of design alternatives, automate routine tasks, and help engineers move from concept to implementation more quickly.

This shifts the designer’s role from performing every task manually to guiding the engineering intent: framing the problem, selecting meaningful alternatives, interpreting results, and deciding what is viable for production.

Human expertise remains essential in judging whether a design satisfies performance, safety, manufacturability, cost, and regulatory constraints. By pairing AI-assisted development with physics-based modeling and simulation, teams can explore and verify more concepts to find solutions that best meet their goals and requirements.

How is AI applied within Model-Based Design for EV battery systems, including thermal management, ageing prediction, and system-level optimization?

AI-augmented Model-Based Design helps EV teams by combining physics-based models, simulation, and AI-driven insights to evaluate battery performance before hardware is built.

AI can help engineers improve battery thermal management, state-of-charge and state-of-health estimation, ageing prediction, fault detection, and remaining useful life analysis. Model-Based Design provides the framework to simulate, test, and validate these capabilities. For example, AI-based reduced-order models can accelerate battery and thermal system simulations by replacing computationally expensive high-fidelity component models with fast surrogate models, enabling faster design iterations and real-time testing. AI-powered virtual sensors can also estimate difficult-to-measure battery states, improving battery management system (BMS) performance without additional hardware.

At the vehicle level, multi-domain simulation helps engineers optimize batteries alongside motors, power electronics, thermal systems, and controls. This allows engineers to evaluate trade-offs in range, charging performance, safety, durability, and energy efficiency before physical prototypes are built.

The result is better-informed engineering decisions and greater confidence in battery and vehicle performance while maintaining engineering rigor through simulation and validation.

What role do you see for Model-Based Design in managing the complexity of software-defined vehicles?

As vehicles become increasingly software-defined, a model-based approach provides the foundation for managing complexity at scale.

Software-defined vehicles are introducing unprecedented levels of software complexity. Functionality is increasingly distributed across centralized computing platforms, zonal controllers, sensors, actuators, and connected services. Managing these interconnected systems requires a disciplined development approach that spans architecture, software, controls, and verification.

A common challenge among vehicle manufacturers and suppliers is that software complexity is growing faster than team size. This creates a need for clearer alignment between requirements, architecture, implementation, and testing.

Model-Based Design provides a framework for managing this complexity through architecture-driven development, requirements traceability, simulation-based verification, and automated code generation. As agentic AI capabilities mature, the opportunity is to close the loop further by helping engineers move from requirements to architecture, simulation, analysis, and test within a connected engineering environment.

A model-based approach also supports collaboration across teams, reuse across vehicle platforms, and shift-left verification that identifies issues earlier as vehicle software becomes more interconnected and continuously updated.

What role do MATLAB Copilot, Simulink Copilot, Polyspace® Copilot, and the MATLAB and Simulink Agentic Toolkits play in helping automotive engineers move faster from design and implementation to testing and verification without compromising engineering rigor or safety?

MATLAB Copilot, Simulink Copilot, and Polyspace Copilot serve as interactive AI assistants, while the MATLAB and Simulink Agentic Toolkits enable AI agents to participate in engineering workflows within trusted Model-Based Design environments. Together, they bring AI-driven automation into established engineering processes, using the same methods and tools engineers rely on to verify, validate, and review AI-produced work.

For automotive engineers, these capabilities can assist with activities such as understanding legacy code, navigating complex models, debugging software, analyzing code-quality findings, generating tests, and validating system behavior. Agentic workflows can help automate tasks such as executing simulations, generating and running tests, analyzing software quality issues, and supporting troubleshoot-fix-verify workflows.

By connecting AI directly to engineering artifacts, including requirements, models, software, tests, and verification results, teams can reduce time spent on repetitive tasks and focus more on evaluating design trade-offs, system optimization opportunities, and validation outcomes. Engineers remain in control, with all AI-assisted outputs subject to the same simulation, testing, verification, traceability, and review processes required for automotive systems.

These capabilities can help improve engineering productivity while maintaining the rigor, quality, safety, and reliability required for modern automotive engineering.

How does AI help automotive engineers train autonomous driving algorithms and deploy them as reliable embedded intelligence?

Autonomous driving systems rely on AI for perception, sensor fusion, object detection, path planning, and decision-making. Engineers train these algorithms using both real-world and simulated data, while simulation enables evaluation across thousands of diverse and safety-critical driving scenarios that would be difficult, expensive, or unsafe to reproduce on public roads. AI is also increasingly supporting engineers through agentic workflows that orchestrate scenario generation, testing, validation, and design exploration across engineering tools, helping teams develop and refine autonomous driving functions more efficiently.

Reliability comes from combining AI with rigorous engineering workflows. Autonomous driving algorithms are validated through simulation, software-in-the-loop, hardware-in-the-loop, and system-level testing before deployment. Once validated, AI models are optimized for embedded vehicle hardware using techniques such as quantization, pruning, and model compression to meet real-time operation, memory, and power constraints. By connecting requirements, models, simulations, tests, and deployment artifacts through a digital thread, automotive teams can continuously improve AI systems while maintaining the traceability, verification, and safety required for production vehicles.

Closing thought

“The goal is not simply to develop software faster. It is to help engineers build better products through engineering rigor, system-level insight, and AI-enabled productivity.”

Across the automotive industry, we’re seeing a shift from isolated engineering domains to integrated, software-defined systems. At the same time, AI is becoming a practical engineering tool rather than a futuristic concept.

The organizations creating the greatest impact are those that combine AI, data, simulation, and engineering expertise within a unified development workflow. The next wave of productivity will come from workflow-aware agentic AI skills that can inspect models, run simulations, analyze outcomes, and troubleshoot issues. These agents will help engineers close the loop between requirements, design, verification, and continuous refinement.

About Vijayalayan R

Vijayalayan R leads Automotive Industry and Location Field Application Engineering at MathWorks India, helping customers adopt Model-Based Design for next-generation automotive programs spanning electrification, AI, and software-defined vehicles. He also contributes to the automotive ecosystem as secretary of the SAE India Bengaluru Section and volunteers as a STEM mentor with Katalyst India. With over 24 years of experience, he holds a master’s degree in control and instrumentation from IIT Madras.

Also read: Exploring Battery Modelling and Simulation using Data and AI