In my company, the role of “AI Software Engineers” primarily involves building applications and workflows utilizing large language models (LLMs) rather than training or creating them. Their work encompasses developing product features and internal tools, heavily relying on LLM inference APIs and vector databases. Much of their effort goes into writing glue code to manage LLM context, querying data, and integrating with other systems, alongside creating front-end interfaces for their applications. Recently, they’ve begun to explore training LLMs using managed services like AWS SageMaker, but this is still in the early stages. The expectations of the title “AI Engineer/SWE” can vary significantly based on an organization’s technical sophistication; for instance, roles at OpenAI are likely to involve fundamental contributions, whereas elsewhere, they may not be as advanced.
Source link
Share
Read more