7 votes

Most AI job listings ask for computer scientists with a PhD. But if you want to move into AI, there are options for experienced IT pros who can deliver on an ML project

2 comments

  1. clerical_terrors
    Link
    Fondly enough our professors in my AI degree are now telling us to be careful when looking on Linkedin for companies to do a research internship or a thesis project at, because the vast majority...

    Fondly enough our professors in my AI degree are now telling us to be careful when looking on Linkedin for companies to do a research internship or a thesis project at, because the vast majority of specs they ask for are too simple to make for worthwhile academic research.

    3 votes
  2. onyxleopard
    Link
    I feel like there are some semantic shenanigans going on here. You need both scientists and engineers to move any ML application from development into production. There are very, very few...

    I feel like there are some semantic shenanigans going on here. You need both scientists and engineers to move any ML application from development into production. There are very, very few engineers who know enough of the science, and even fewer scientists who know enough about deploying commercial software. Making a distinction between ML and AI is not helpful. AI has lost all meaning in practice due to abuse. Deep learning is still ML just as much as simple regression models. Is a program making decisions based on real-world data? Congrats! That’s a machine learning application. Did you implement the algorithm that makes the decision? Congrats! You’re an ML research scientist. Did you gather, clean, and transform the raw data? Congrats you’re a data engineer. Did you experiment with various models and parameters to select an appropriate model and configuration? Congrats you’re a data scientist. Did you put the final model into production? Congrats, you’re a devops engineer. Did you do more than one of the above? Oops! Your org doesn’t know what it’s doing (and you’re probably not being paid enough or you probably didn’t do a good job or you took much longer than it should have).

    3 votes