On August 12, 2025, the Global Center for Student Innovation and Entrepreneurship of Peking University held an MBA seminar focused on future energy and artificial intelligence.
As a representative of a PKU-alumni enterprise and a lecturer of the Peking University Innovation and Entrepreneurship Training Camp, Ting Yang, Vice President of Sointech Group and CEO of SOIN AI, and CTO Tengyu Hu were invited to teach the course "AI Empowers Manufacturing: A Scenario Revolution from Intelligent Design Assistance to Knowledge Services," systematically revealing SOIN AI's forward-looking industry deployment strategy and core technical achievements in Agentic AI.
At the event, Ms. Yan Han, General Manager of Enterprise Services at the Peking University Entrepreneurship Training Camp, reviewed the camp's development and achievements, sharing the platform's successful experience in project incubation, capital matching and international expansion; the head of Peking University's Technology Development Department focused on policies and typical cases of scientific achievement transformation, showcasing PKU's active exploration of industrializing research results.
Core Views from PKU AI Class Lecturers
Lecturer Ting Yang: Defining a New Industrial Paradigm for Agentic AI
CEO Ting Yang deeply explained SOIN AI's strategic blueprint for defining a new industrial paradigm of Agentic AI. She pointed out that enterprise-grade agents are becoming the core form of AI empowering the real economy. Through a three-dimensional strategy of "technology — scenario — ecosystem," SOIN AI is committed to building subscribable, evolvable industry agent solutions:
Strategic anchor: with manufacturing PLM (product lifecycle management) and education publishing knowledge services as dual engines, driving industry's deep leap from informatization to intelligence;
Business practice: focusing on "must-have cost-reduction and efficiency scenarios," restructuring core enterprise business processes through agent workflows, making knowledge assets programmable and service-oriented;
Ecosystem layout: in 2025, officially launching the "AWE Engine Subscription Partner Program," joining developers to build a flourishing agent ecosystem.
Lecturer Tengyu Hu: The AWE Engine Tackles Manufacturing's Core Scenarios
CTO Tengyu Hu revealed for the first time in depth the breakthrough applications of the proprietary Agentic Workflow Engine (AWE) in manufacturing:
Intelligent design revolution:
PLM AI-assisted design system: embedding industry knowledge bases and design standards into dynamic workflows, significantly improving design efficiency while optimizing manual verification costs;
Electrical design automation: AI-assisted BOM generation and bid quotation, solving the data fragmentation and slow response pain points of traditional processes, reducing data error rates by more than 70%.
AWE engine core technical moats:
Dynamic workflow orchestration: supporting modular programming of business processes, flexibly adapting to manufacturing's complex decision chains;
Multimodal data processing: seamlessly fusing text, drawings, parameter tables and other industrial data to build industry knowledge graphs;
Cloud-native microservice architecture: ensuring enterprise-grade reliability and elastic scaling.
From Peking University to Industry: The Agent Ecosystem Accelerates Deployment
SOIN AI's industry agent solutions have already signed cooperation agreements with multiple KA customers in manufacturing and education publishing.
Going forward, with the AWE engine at its core, the company will continue to build vertical-industry agent technology standards, advancing AI from an "efficiency tool" to an "intelligent foundation" that drives industrial leaps, providing solid technical support for Intelligent Manufacturing in China.