This repository serves as the course material for "Generative AI Application Systems and Engineering (AIASE 2026)" at NCKU CSIE. It targets students and engineers seeking to master the end-to-end development lifecycle of Generative AI applications, from requirement analysis and system design to implementation and deployment, culminating in a functional Generative AI web service. The course provides a comprehensive curriculum covering modern software and AI engineering practices.

How It Works

The course adopts a project-centric approach, guiding students through the Software Development Life Cycle (SDLC) applied to Generative AI. Core concepts include front-end/back-end development, data flow orchestration, CI/CD pipelines, MLOps/LLMOps, and cloud-native architectures on AWS. Emphasis is placed on practical implementation of advanced topics like LLM fine-tuning, agentic workflows (MCP/ADK/Agent-to-Agent), prompt engineering, and mitigating LLM security concerns.

Quick Start & Requirements

• Prerequisites: Basic programming proficiency, web technologies (HTML/CSS/JS), GitHub usage experience, and a foundational understanding of cloud services are required. Self-directed troubleshooting skills are essential.
• Resources: Links to course materials, LLMOps Overview, SDLC Overview, and documentation for MCP, LLMOps, and Ray are provided.
• Community: A Discord server is available for course discussions: Taica 生成式 AI 課程 (Note: Actual Discord link not provided in README).

Highlighted Details

• Covers advanced Agentic Engineering concepts, including Memory Engineering, ReAct/Reflection/Multi-Agent architectures, and Spec-Driven Development (SDD).
• Integrates MLOps/LLMOps practices for managing AI model lifecycles within application development.
• Explores practical aspects of LLM deployment, including AWS cloud environments, distributed AI infrastructure, and token economics for cost control.
• Addresses critical LLM challenges such as prompt optimization, multi-turn dialogue design, hallucination reduction, and AI security.

Maintenance & Community

This repository functions as an educational resource for a specific academic course. Maintenance is likely tied to the academic calendar and instructor updates. Community interaction is facilitated via a Discord server.

Licensing & Compatibility

No specific software license is mentioned for the repository's content. As an academic course repository, its primary use is educational, and direct integration into commercial or closed-source projects may require clarification from the instructor.

Limitations & Caveats

This repository contains course materials and assignments, not a production-ready software framework. The content is structured for learning and may not represent the most current or stable implementations of the technologies discussed. Specific versions of tools or libraries used in assignments may not be explicitly detailed, requiring students to adapt based on course guidance.