The short answer for the title is of course: different. This would be a very short talk otherwise. The experience of 10+ LLM based application development projects were distilled into this talk. It mainly focuses on differences compared to other kinds of projects and advice on what to expect to change when moving from proof-of-concept to production. As we have seen multiple times this shift requires proper expectation management, hence foresight is much needed, therefore this talk came to be.
Sandor Kazi has worked as a data solution architect for EPAM the part 2+ years, working with data and machine learning his entire professional career it was natural to him to pursue "generative AI" application projects from an early stage of the area. He has designed and lead the development of multiple data platforms and applications for various use-cases for EPAM's customers on various cloud pla...