Angielski
Software Engineering professionals now face challenges in incorporating GAI into the products and programs they are developing. At this point, the knowledge about developing AI-based software is mostly based on classical AI, i.e., non-generative ML systems. Developers know how to use machine learning and, to some extent, how to include it in production systems. Engineering Generative-AI Based Software takes software development to the next level by using generative AI instead. Readers learn how to use text, image and audio models as part of larger software systems. The book discusses both the process of developing such software and the architectures for this kind of software, combining theory with practice. Generative AI software is gaining popularity thanks to such models as GPT-4 or Llama. More and more products use them as part of their feature portfolio, but this software is often limited to web applications or recommendation systems. Author Miroslav Staron shows readers how to tackle the challenges of professionally engineering generative AI-based systems. The book starts by reviewing the most relevant models and technologies in this area, both theoretically and practically. Once readers know the technologies, the book goes into details of software engineering practices for such systems, e.g., eliciting functional and non-functional requirements specific to generative AI, various architectural styles and tactics for such systems, and different programming platforms. The book also shows how to create robust licensing models and the technology to support them. Finally, readers learn how to manage data, both during the training and also when generating new data, as well as how to use the generated data and user feedback to constantly evolve generative AI-based software.
