This is a broad question.<p>AI/ML needs a math-heavy background and also a lot of domain knowledge to make any sense.<p>I think the best path for software engineers to go into data is via data engineering or MLOps or whatever it is called tomorrow. This is relatively close to software engineering in many cases, however it's becoming more and more about writing YAML files, too.<p>What these field could really benefit from though is some software engineering best practices such as testing, linting, formatting and so on.
It's a job like any other. And it's not outside of SWE, it's part of it; this is a bit like asking how someone moved from SWE to being a backend developer.
Switched to an AI team 8 months ago. Ended up doing infra and api integration for llm tuning which is a nice way to start ramping up to the field. With every integration learned new things: Lora/pets, spin up servers to host the models for inference/tuning. It's a lot of learn on the spot compared to concepts from SWE life.
My personal experience comes from the following, I was a full stack person who eventually showed a high level of talent for frontend and made that my main expertise. However, with the demand for AI/ML/Analytics being big in my area I have had to do a hard transition. I have followed AI/ML/Analytics for some time and I have a foundational knowledge of the concepts behind it. However, it's taking theory/concepts to actual real world human interfacing applications that meet with what client's expectations that is the big hurdle. Learning to rein in the expectations is a skill that I feel is worth mastering at the same time.
I don't really hear from them anymore, I am guessing they have their hands full and can imagine the pressure they might be on? could be completely wrong