Professor-level Senior Engineer | Industrial AI Expert | LLM Model Evaluator
Specialized in LLM evaluation and adaptation for industrial scenarios, transforming 37 years of ship engineering experience into AI model safety constraints and testing frameworks
37 years deep experience in ship industry, from traditional ship design to unmanned vessel system integration, and then to industrial AI fusion transformation. International vision (Singapore, USA, Europe work experience), leading multiple national key projects, with 2 national invention patents and 15+ utility model patents.
2025 Career Transformation Breakthrough: At 59 years old, with zero coding experience, completed full-stack development and AI Agent integration independently within 1 month with Claude's assistance, achieving deep integration of industrial knowledge and AI technology, exploring innovative paths for intelligent upgrading of traditional industries.
Mastered Python and full-stack development from zero basis, implementing AI Agent applications in ship engineering
2 national invention patents, 15+ utility model patents, independent R&D of high-speed propulsion device series products
Work experience in USA, Singapore, Europe, leading China-Russia cooperation and national 863 major projects
Founded high-tech enterprise in 2012, awarded "Huanghe Talent" leading talent and multiple honors
Applying 37 years of industrial engineering experience to LLM evaluation and safety alignment, systematically comparing Deepseek, Claude and other models in industrial scenarios
Design test sets for industrial scenarios, evaluating LLM performance in ship design, unmanned vessel systems, safety regulations, etc. Including multi-dimensional scoring system (accuracy, safety, engineering usability) and systematic Prompt optimization methods.
Three-layer architecture: RAG knowledge retrieval + tool calling + rule validation, ensuring AI outputs comply with engineering standards and safety requirements. Validated feasibility in propeller design, ship resistance calculation scenarios.
Flexible collaboration approaches to assist model landing and safety alignment in vertical domains.
Real industrial scenario LLM evaluation cases, including detailed Prompt design, model output comparison, scoring standards and improvement suggestions
| Dimension | Weight |
| β Technical Accuracy | 35% |
| β Engineering Usability | 30% |
| β Safety Compliance | 20% |
| β Response Quality | 15% |
Multi-dimensional scoring system designed based on industrial engineering standards
Systematic comparison of mainstream models in engineering calculation scenarios
Complete portfolio contains 10+ real industrial scenarios detailed evaluation reports
Full PDF version available for job applications or collaboration discussions, demonstrating systematic model evaluation methodology and practical experience
Request Full Portfolio βLooking forward to collaborating with you on exploring the integration of ship engineering and AI technology