retooling from data engineering into the agent stuff, is anyone actually hiring for this or is it all hype

spent about 8 years doing pipelines, airflow, dbt, the usual. the market for that dried up and everyone keeps telling me to pivot into agents and llm stuff. so heres my genuine question, when a company posts an ai engineer role, are they hiring people who wire up agents and mcp servers, or are they really just hiring ml phds and calling it ai. because those feel like two totally different jobs and the listings never say which. anyone made this exact jump recently and can tell me what they actually wanted in the interview?