MathWorks is showcasing its latest advancements in AI for safety-critical systems at NeurIPS 2025 in San Diego, highlighting how MATLAB and Simulink empower engineers and scientists to design, verify, and deploy AI-driven solutions in high-stakes domains such as aerospace, automotive, robotics, and electrification. The company is presenting at Booth #732 during the world’s leading conference on machine learning and computational neuroscience, which draws thousands of researchers, data scientists, and industry leaders from over 70 countries. Lucas Garcia, AI Product Manager at MathWorks, emphasized the importance of bridging the gap between academic AI innovation and real-world engineering reliability. “NeurIPS is where cutting-edge AI research meets real-world impact,” Garcia said. “We’re demonstrating how Model-Based Design with MATLAB and Simulink supports the full journey from prototype to production—enabling safe, verifiable, and deployable AI systems in complex engineering environments.” At the event, Garcia will deliver a presentation on Tuesday, December 2, at 3:00 PM PDT in Exhibit Hall A, B, titled “From Research to Trust: Building Certifiable AI for Safety-Critical Systems.” The talk will address the challenges of deploying AI in regulated industries, where predictability, explainability, and robustness are essential. He will discuss how formal verification, model explainability, and runtime monitoring can transform black-box AI models into systems that meet rigorous safety standards. The session will feature practical examples using models developed in PyTorch, ONNX, and MATLAB. MathWorks will also present three interactive demonstrations at its booth: AI-Based Virtual Sensor for Battery State-of-Charge Estimation – This demo shows how AI can be combined with physics-based models to create virtual sensors that accurately estimate battery state-of-charge. Using MATLAB and Simulink, engineers can simulate battery behavior, compress models for efficient deployment, and generate production-ready code—critical for electric vehicles and energy storage systems. RTM Pose Estimation Deployed to Embedded GPUs – This demonstration highlights real-time pose estimation for robotics and autonomous systems. It illustrates how models from PyTorch or ONNX can be imported into MATLAB, optimized for performance, and deployed to NVIDIA Jetson AGX Orin hardware. The workflow supports rapid prototyping and efficient deployment in resource-constrained environments. MCP with MATLAB – MathWorks will show how a large language model (LLM) client can integrate with the MATLAB MCP Core Server to automate data cleaning, generate visualizations, and build simple machine learning models. The demo is designed for quick iteration and reuse, streamlining common data analysis tasks. MathWorks, founded in 1984, is a global leader in mathematical computing software. Its products, MATLAB and Simulink, are widely used in engineering, research, and education across industries including automotive, aerospace, electronics, and industrial automation. With over 6,500 employees in 34 offices worldwide, MathWorks continues to support innovation through tools that accelerate development and ensure system reliability. For more information, visit mathworks.com.