Agents are amazing. Harnesses are cool. But the fundamental role of a data scientist is not to use a generalist model in an existing workflow; it's a completely different field.
AI engineering is the body of the vehicle, whereas the actual brain/engine behind it is the data scientist's playground.
I feel like I am not alone in this realisation that my role somehow got silently morphed into that of an AI engineer, with the engine's development becoming a complete afterthought. Based on industry requirements and ongoing research, most of the work has quietly shifted from building the engine to refining the body around it.
Economically, this makes sense, as working with LLMs or other Deep Learning models is a capital-intensive task that not everyone can afford, but the fact that very little of a role's identity is preserved is concerning.
Most of the time, when I speak to data scientists, the core reply I get is that they are fine-tuning models to preserve their "muscles". But fine-tuning is a very small part of a data scientist's role; heck, after a point, it's not even the most important part. Fine-tuning is a tool. Understanding, I believe, should be the fundamental block of the role.
Realising that there are things other than "transformers" and finding where they fit into the picture. And don't even get me started on the lack of understanding of how important the data is for their systems.
A data scientist's primary role is not the model itself. It's about developing the model, the data quality at hand, the appropriate problem framing, efficiency concerns, architectural literacy, evaluation design, and error analysis. Amid the AI hype, many have overlooked that much of their role is static and not considered important.
AI engineering is an amazing field. The folks who love doing amazing things with the models always inspire me. But somehow, the same attention and respect are no longer paid to the foundational, scientific side of data and modeling in the current industry. I realise it's not always black and white, but it's kind of interesting how the grey is slowly becoming darker by the day.
Do you feel the same way? Or is it just my own internal crisis bells ringing unnecessarily?
For those of you who have recognized this shift, how are you handling your careers? Are you leaning into the engineering/systems side and abandoning traditional model development? Or have you found niche roles/companies that still value the fundamental data scientist role (data quality, architectural literacy, statistical rigor)? I'd love to hear how you are adapting
[link] [comments]
