(ML Engineer is in my title)

My impression is that there's more demand for ML-oriented roles than for engineers in general, but it's still a poor job market compared to pre-2020. I hear from recruiters occasionally but mostly just dubious startups. The companies I interact with (mid-sized non-tech) are reluctant to hire anyone for any reason, even though management is full of ideas for new AI products.

I don't think I can offer advice on whether new grads should focus on it or not. In the short term it might be a really good move, especially if you're unusually talented, but if everyone is doing that and then the hype dies down a bit you might end up in a tough spot.

I know someone who graduated college recently with a focus on ML. They have been looking for a job but can’t find anything.

I told them to go more general skill or join/start a company but they’d rather just not broaden their horizon /shrug

Another friend who was experienced ML engineer is now a mobile engineer.

1. All the changes in AI from the past couple of years have only increased demand for MLEs. The job profile kind of widened too, - MLE has always been a subset of SWE but even more so if you're working with LLMs. So some MLE jobs will still require a lot of theory, some won't, but they all require some understanding of AI/ML. You kind of need to know what you want to work on instead of targeting specific job titles (which can be hard to navigate if you're early in your career and have no experience).

2. Overall market dynamics still apply - the increase from 1. is relative to the market i.e. if the market has been tough for other roles like frontend for example, it has been less tough for MLEs because they work on the thing that's exciting at the moment.

3. Should new grads focus on it? New grads will likely have a difficult time regardless of what they decide to do - maybe in a couple of years there won't be excitement around AI/ML, maybe there will be more, maybe something else will be hot. The overall market almost certainly be different, and that could be for the better or worse.

Hard to predict the future. Whatever you decide to do early in your career, know that you're not married to it forever. Develop general skills that can be applied in many areas.

Does it make a difference if they have had internships or have open source contributions or a strong github profile with non trivial projects?

Incorporating LLMs into software is a software dev job from what I have experienced. You don't need to know anything other than how to design and integrate APIs. So, the real answer to your question is dependent on what companies are actually "building" something versus which are simply slapping together a bunch of things (I am not against either, if my phrasing suggested otherwise!). A lot of companies and research groups are emerging where you do need ML theory and hence ML engineers. Take for example PPFL (Privacy Preserving Federated Learning) or similar fields.

Even ML engineer jobs are hard to come by. Not sure how many job postings are real.

In my current experience, in applying to 300 jobs, I am qualified for. Leads to exactly 1 interview.

Very tough, unusual IT market.