Data science career path shifts from ML training to building LLM apps and agents

mdancho84 · x · 2026-09-12

The author recommends Mathematics of Machine Learning, which ties linear algebra, calculus, probability and optimization directly to why ML works. But he argues learning ML is no longer enough: companies now need people who can build LLM apps, agents, RAG systems, APIs, databases, tools, evals and automations — meaning the classic Python→SQL→Stats→ML roadmap is giving way to AI engineering skills.

Original post →

More from coding & agent

coding & agent channel →