I’m seeing a lot of anxiety among graduates. Since Generative AI (ChatGPT, Gemini) can now write SQL queries, clean datasets, and even generate Matplotlib visualizations, what is left for the junior data scientist? Should we be focusing more on domain expertise or advanced mathematical modeling to stay competitive?
3 answers
The role isn't disappearing, but it is evolving rapidly. Entry-level tasks that used to take three days now take three hours with Generative AI (ChatGPT, Gemini). This means junior scientists are expected to perform at what was previously a mid-level capacity. You need to focus on "problem framing" and "data storytelling." Anyone can generate a plot, but explaining why a specific outlier is affecting the business logic is where the human value remains. Don't just learn to code; learn to ask the right business questions.
Do you think that the bar for technical interviews will be raised to include more complex architecture instead of just coding?
I think juniors should focus on MLOps. The AI can write the model code, but it can't monitor the model in production or handle data drift yet.
Spot on, Brenda. Operationalizing AI is much harder than just generating a script, and that’s a huge opportunity for new professionals.
Absolutely, Edward. We are already seeing interviews shift away from LeetCode-style questions toward system design and how one manages AI-driven workflows effectively.