I'm learning ML right now, but I think I actually want to become an AI Engineer. What should I realistically learn?
Post:
I'm a 3rd-year B.Tech CSE (AI) student wasted 2 years and I'm currently learning Machine Learning through a fairly comprehensive ML course.
Initially, I thought I wanted to go into Data Science/ML, but after learning more about the industry, I think I'd rather pursue AI Engineering — building applications and systems using existing models rather than primarily focusing on developing ML models from scratch.
This is where I'm confused.
what i thought i will do:
ML fundamentals → Deep Learning → NLP/Transformers → LLMs/GenAI → RAG → AI Agents → FastAPI/deployment → projects
DID prerequisites for ml almost took 2 months and was learning ML since 1.5 month
But I'm not sure if this is actually the right path.
For example, do I need to become really strong at traditional ML before moving into AI Engineering? Should I properly learn Deep Learning and PyTorch, or can I move relatively quickly toward LLM/GenAI development after getting the necessary ML fundamentals?
And how important are things like:
- FastAPI
- PyTorch
- Transformers
- RAG
- Vector databases
- LangChain/LangGraph
- AI agents
- Docker/cloud/deployment
- SQL/databases
- system design
for someone trying to get their first AI/ML Engineering internship or job?
So if you're currently working as an AI Engineer / ML Engineer / Applied AI Engineer, I'd really appreciate your honest advice:
If you were a 3rd-year student starting from my position today, what would you learn, what would you skip, and what would you actually build to become employable?
I'm especially interested in hearing from people who are actually working in the field rather than just following online roadmaps.
Thanks!