Disruptive AI in Product Design

Interview with Théophile Allard, the CTO of Neural Concept

Disruptive AI is revolutionizing product design, particularly in 3D modeling and engineering. The platform developed by Neural Concept enhances efficiency and innovation across various industries, including automotive and aerospace. By leveraging AI and cloud technology, it accelerates development and optimizes performance, ultimately transforming product development processes.

Highlights

00:17 AI engineering is a crucial reality today that is reshaping multiple industries. Companies like New Concept are pioneering in applying deep learning technologies for innovative product development.

• Neural Concept, a Swiss startup, has made significant strides in AI by developing software for 3D AI models in various sectors, notably automotive and aerospace.

• The technology developed by New Concept is versatile and applicable across multiple industries, including consumer electronics, aerodynamics, and structural mechanics.

• Partnership with Microsoft enhances New Concept's capabilities by leveraging Azure's scalable cloud infrastructure, enabling them to manage AI workloads efficiently.

04:05 The development of AI-powered engineering solutions has evolved significantly, particularly through the creation of a modular platform that empowers engineers in product development workflows. This platform leverages advanced predictive models and enables customization for various industries.

• The early achievements of the company included collaborations with major clients like Airbus, showcasing their technology at top machine learning conferences. These collaborations helped establish their credibility in the industry.

• The transition from a fully integrated platform to a modular one reflects the need for flexibility in industrial applications. This shift allows for tailored solutions that accommodate specific user needs and expertise.

• The importance of interactive development environments is highlighted, enabling engineers to combine various models and create custom workflows. This adaptability is crucial for successful integration of AI in diverse engineering applications.

08:10 The advent of large language models significantly accelerates innovation and reduces development time for engineers, enabling faster and more efficient tailored solutions. This transformation is driven by faster execution and simpler integration of AI models.

• Key performance indicators (KPIs) play a crucial role in measuring the impact of AI models on development speed and product performance. They help in assessing overall efficiency improvements.

• Integrating AI models into operational digital twins enhances the efficiency of systems by providing real-time responses and continuous learning through reinforcement learning capabilities.

• The example of a team attempting to break the wind-powered speed record illustrates how simulation-based digital twins can optimize real-world performance through advanced modeling techniques.

12:14 The collaboration between technology and sports, particularly in sailing, showcases how advanced simulations can optimize performance and safety. This is evident through the use of digital twins and predictive models.

• Digital twins, powered by Neural Concept, allow pilots to train effectively before test runs, enhancing their control over the boat during high-speed conditions.

• Innovations in sailing, such as specialized foils designed for extreme speeds, address unique physical challenges like cavitation, showcasing the need for advanced engineering solutions.

• In the automotive sector, technology is helping manufacturers like Subaru improve aerodynamics, significantly reducing carbon footprints and enhancing electric vehicle battery ranges.

16:17 Cutting-edge process automation facilitates seamless integration of various CA and Cat tools, allowing customers to enhance their production applications and streamline processes effectively. This innovative approach also accommodates custom models, ensuring adaptability in technology integration.

• The integration of multiple AI models enables real-time communication for optimizing different components, which enhances the overall system performance and efficiency. This collaboration is essential for complex projects.

• Challenges in transforming IT environments are addressed by creating scalable and elastic infrastructures that support advanced technologies, promoting a smooth transition for engineering organizations. This collaboration with Microsoft is crucial.

• The emphasis on high-performance computing environments and cloud power is vital for harnessing the full potential of AI technologies. This ensures that infrastructures are adequately prepared for new innovations.

20:20 Data organization is crucial for model performance, and companies vary in their ability to manage and structure this data effectively. Mature organizations can leverage extensive datasets, while others face significant challenges.

• Some companies, like Formula 1 teams, excel in data management with vast datasets that enhance model training. Their structured approach enables high performance in simulations and predictions.

• Active Learning is a key feature that allows models to initiate additional training based on performance feedback. This real-time adjustment improves model accuracy and efficiency.

• Security in cloud environments is essential, particularly with sensitive data. Collaborations with companies like Microsoft focus on ensuring robust security measures and compliance certifications.

24:24 The company aims to transform the industry by enhancing product development through virtualization, addressing societal challenges like energy efficiency. This innovation is expected to accelerate technological advancements significantly.

• The long-term goal includes doubling or even multiplying the speed of technology innovation, leading to better energy efficiency in industrial systems and products.

• Upcoming Neural Concept version 4.3 will improve data organization and tailored workflows, enhancing user experiences in product development.

• Integrating with Rescale for on-demand simulations will provide users with seamless thermal management and external aerodynamics applications, minimizing IT efforts for customers.