Is the ds/ml slowly being morphed into an AI engineer? [D]
Data scientists are becoming AI engineers, but who's building the engine?
A viral Reddit post from user The-Silvervein has sparked a heated debate in the AI community about the evolving identity of data scientists. The author argues that the role of a data scientist has silently morphed into that of an AI engineer, with the core focus shifting from developing models to refining workflows around existing generalist models like LLMs. They emphasize that while AI engineering is valuable—building the 'body' of the vehicle—the 'brain' or engine development (data quality, problem framing, architectural literacy, evaluation design) is becoming an afterthought. The post notes that fine-tuning, often cited as a way to preserve data science skills, is just one tool and not the core of the role.
Economically, this shift makes sense: training deep learning models is capital-intensive, so most companies now leverage pre-trained models. But the author warns this devalues foundational data science skills, such as statistical rigor and error analysis. The post has resonated widely, with many commenters sharing similar career dilemmas—whether to lean into AI engineering or seek niche roles that still value traditional model development. This reflects a broader industry trend where the hype around LLMs and agents overshadows the scientific side of data and modeling, leaving data scientists to reassess their professional paths.
- Data scientists' roles are shifting from model development to AI engineering with LLMs and fine-tuning
- Core data science skills like data quality, statistical rigor, and architectural literacy are being devalued
- Economic pressures make capital-intensive model training less accessible, pushing professionals toward engineering roles
Why It Matters
This signals a fundamental industry shift that data professionals must navigate to stay relevant.