With the rise of LLM's there is definitely a new skillset needed in the bag of engineers - one who know how to align/instruct LLM's towards an objective. This n

With the rise of LLM's there is definitely a new skillset needed in the bag of engineers - one who know how to align/instruct LLM's towards an objective. This new role is essentially a mix of standard programming logic + ability to instruct LLM's using language (natural language programming). That's been the primary difference I have seen in this emergent field of designing agents. Most agents need some form of formal instructions tied with designing prompts - which are a mix between formal programming and natural language. This needs a slightly different skill set from the traditional backend engineer or services engineer. This is not a ML or data role either. Rather than just a prompt engineer, the case for which could be made like a specialist designer - this role maybe more akin to a full-stack developer role who knows how to align LLM's or Diffusion models towards a multi-step goal.