Question to people in the field. How easy is it to break into AI?
I'm a 23 year old junior engineer doing standard backend infra. Ive gotten interested in AI safety and I'm wondering what the easiest way for a junior backend engineer like me to change to a role where they are working on deep learning models. Interested in eventually landing a research engineer job like this: https://boards.greenhouse.io/deepmind/jobs/5611918 https://jobs.netflix.com/jobs/310850697 rather than a role as a research scientist so hopefully I don't have to go back to school
Researcher? Get a PhD. ML application engineer, a few boot camp you are good lol.
Yah I'm interested in roles like this: https://jobs.netflix.com/jobs/310850697 https://boards.greenhouse.io/deepmind/jobs/5611918 So research engineer rather than scientist. I've taken a few ml related classes in college and done some self studying
Anything with “research” is generally phd level candidates.
False - why has advice on blind gotten so bad. It’s worse than Reddit.
@block is right. That’s true for 99% of cases. Ofc there are exceptions
I'd say it's not very easy but you can do it! In my opinion, there are two good paths: 1. You can find a backend engineering role in an AI-related company, work there for some time, learn as much as possible about deep learning, and then eventually you will be able to get a research engineer role (most likely, at a different company). 2. Start building personal projects, play with different ML models/services, and just try to apply for research engineering positions.
Appreciate it! I'll look into building some personal projects and shoot my shot at a transfer to a ml research team at my current role
I’ve studied both AI/ML and software engineering. While they’re work great together there’s not a lot from engineering that you can take to ML. Most of the ML is maths-stats. I personally couldn’t get any jobs in the ML field(this was 5 years ago) because all research positions need PhD or years of experience. I didn’t get any interviews. I then shifters my career goal towards engineering. I later discovered that I can probably stay close to ML which still being an engineer with ML engineer roles. I gave it a try and personally found ML engineer little boring because the engineering aspect of it is very easy but a lot of trial and error. You’re working for research scientists who are not always great programmers and my job was mostly to productionize their code.
Also studied both and have worked in ML my whole career, but disagree. You need to know your ML and fundamental stats and have research experience (not necessarily PhD, but papers) to pass the interview, but... once you're in, you immediately realize your MLEs are either nonexistent or overloaded. Depending on your specific situation, MLE work becomes a big chunk of the job if you're at all capable of it. Fairly common for researchers to use platform products to deploy models, or even to own inference services with on-call and all.
Following