I've been staring at job descriptions for AI engineer roles for like three hours now and honestly my head is spinning. I'm currently working as a junior data analyst here in Chicago and I really want to make the jump into actual AI or machine learning development before the end of the year. I’ve been saving up and I have about 2,000 dollars set aside for some intensive online bootcamps or specialized certifications but I really don’t want to waste that money or my limited time on the wrong tech stack.
Everyone says Python is the big one and I get that. I use it a little bit for basic data cleaning at my current job but I've been reading some conflicting stuff lately. One article I found on a tech blog was saying that if you're serious about the engineering side of AI—like actually deploying models at scale or working on self-driving tech—then just knowing Python isn't enough anymore. They were pushing C++ really hard for performance and things like CUDA. But then I saw this long thread on a dev forum about how Julia is going to eventually replace everything for mathematical modeling because it's faster and easier to read.
I'm trying to be really thoughtful about my roadmap because I only have about six months to get proficient enough to pass a technical interview. I'm mostly interested in computer vision or maybe LLM fine-tuning since that's what seems to be blowing up in the industry right now. Here is what I’m looking for in terms of a learning path:
So like... if you had a limited budget and a tight timeline, would you just double down on Python and maybe some SQL or is it really time to start picking up something like C++ or even Rust? I've heard people mention Mojo too but that seems way too new to bet a career on yet. I'm just worried about picking the wrong horse and being stuck...
Stick with Python for now. Be careful about chasing C++ or Rust right now since your timeline is super tight and you might burn out. I would suggest mastering the core libraries first before worrying about performance.
To add to the point above: I was in your shoes just last year! Honestly I almost lost my mind trying to learn C++ because some expert on Reddit told me Python was dying. It was a total waste of my time! I ended up sticking to Python but I realized the real missing piece was understanding the engineering side. I got the DeepLearning.ai Deep Learning Specialization via Coursera and it was literally amazing for my confidence!! I spent like three months just grinding through those labs and the math finally clicked. I also picked up Manning Publications Deep Learning with Python 2nd Edition which is basically my bible now. It's so much better to be a master of one thing than a total newbie at five things... trust me. If you're looking at computer vision, I love using the PyTorch Deep Learning Framework because it just feels so much more intuitive. Stick to what works for the big companies and you'll be fine!!