Westlake University physicist and AI researcher Wu Tailin has launched UniForce AI in January 2026, establishing a dedicated platform for AI-native engineering simulation and control. Building on a decade of research bridging physics and machine learning, Wu aims to transition computational models from academic exploration into industrial-scale engineering infrastructure. Wu’s academic foundation spans a B.S. in Physics from Peking University, a Ph.D. from MIT, and postdoctoral work at Stanford. Since joining Westlake University in 2023 as an assistant professor and head of the AI and Scientific Simulation Discovery Laboratory, he has pioneered neural operator-based surrogate models that accelerate complex system simulations. His work on the AI Physicist framework demonstrates how machine learning can extract fundamental physical laws from observational data. These algorithms have already been deployed in industrial settings, including high-resolution reservoir simulations for Saudi Aramco, and have reduced multi-physics simulation times from hours or weeks down to seconds. Recognizing the limitations of traditional computer-aided engineering tools in handling multi-scale, multi-physics, and strongly coupled systems, Wu founded UniForce AI to develop an AI-native digital base layer for engineering. The company’s initial focus is controllable nuclear fusion, a field Wu describes as the engineering equivalent of AlphaFold due to its extreme computational and control demands. By applying generative simulation, intelligent control, and multi-agent systems, UniForce AI targets the resolution of fusion’s core bottlenecks: high-parameter plasma confinement and multi-scale kinetic modeling. The company has already initiated partnerships with leading commercial fusion ventures and major energy enterprises. The underlying technology leverages latent-space mapping and neural operators to bypass traditional numerical solvers, enabling orders-of-magnitude speedups while maintaining high fidelity and generalization across unseen parameters. Although nuclear fusion serves as the primary validation environment, the methodologies are designed for broad industrial transfer. Potential applications include aerospace propulsion, petrochemical processing, and advanced manufacturing, where similar multi-physics coupling and real-time control challenges persist. Wu positions AI not merely as an additive tool but as foundational engineering infrastructure. UniForce AI currently operates with a lean, interdisciplinary team of eleven researchers spanning physics, machine learning, and systems architecture. Looking ahead, Wu envisions the company delivering a ubiquitous, AI-driven simulation and control ecosystem that continuously learns from physical deployments. Within a decade, the objective is to resolve a paradigm-defining engineering challenge while establishing a new standard for AI-integrated industrial design and operations.