Stephanie Kirmer, a Staff Machine Learning Engineer at DataGrail, discusses how the ML engineering landscape has transformed with the rise of large language models. With nearly a decade of experience in data science and ML, Kirmer brings a unique sociological perspective to analyzing AI's impact on both technology and society.

Kirmer describes the current AI investment environment as a $200 billion bubble, arguing that while LLM technology offers genuine utility, the returns demanded by Silicon Valley investors are unrealistic. She emphasizes that the technology itself isn't to blame for the hype cycle, but rather the unsustainable expectations driving massive capital deployment. Her background in sociology shapes her analysis of institutional dynamics and social inequalities in AI development.

On a practical level, Kirmer notes that LLMs have fundamentally changed her daily workflow through code assistants that help with idea generation, problem-solving approaches, and routine tasks like writing unit tests. However, she maintains that experienced ML engineers still play crucial roles in tackling unique problems. she advocates for AI companies to focus on realistic applications and public education about what LLMs can actually do, rather than perpetuating unrealistic promises that fuel public skepticism.