Under the background of the new engineering education, the cultivation mode for innovation and entrepreneurship talents in universities is crucial for the quality of the talent development. The traditional innovation and entrepreneurship talent cultivation mode faces issues such as unsound talent cultivation system, single structure for talent assessment and evaluation, inability to comprehensively assess the results of talent cultivation. To address these problems, the computer-related majors at Hebei University of Engineering integrate the "learning, training, competition, entrepreneurship" philosophy. A four-level "progressive" ability cultivation mechanism is also incorporated, which includes "learning in class" "practicing in training" "thinking in competition" and "leading in entrepreneurship". This approach is embedded throughout the entire innovation and entrepreneurship talent cultivation mode. The goal is to establish a "progressive integration" mode that merges learning, training, competition, and entrepreneurship. In this vein, innovation and entrepreneurship are embedded throughout the entire talent development process, with goal management spanning the entire professional education cycle. Also, the "1N2N" enterprise project-embedded training system is constructed to improve students' innovation and practical abilities through applications like AI-driven data analysis, personalized learning path design, and intelligent evaluation. In this way, an "AI + dual integration" system for innovation and entrepreneurship talent cultivation is established. Guided by the PDCA cycle theory, the cultivation mode undergoes tracking, feedback, evaluation, and optimization. The quality of talent cultivation is comprehensively managed, and the assessment and evaluation of talent cultivation are realized.