I’ve been browsing the forums for good topics that fit this sort of question, but none have answered it very effectively for me. I’m relatively new to AI development and I am looking for places to better learn AI, but I’m struggling.
I’m looking for someone who’s been through the beginner stage and knows what it’s like, and hopefully could gift some of their wisdom or journey to help us new people get better.
You said none of the existing threads really answered this for you. I think that’s because “get better at AI dev” splits into very different paths: training or fine-tuning models, building apps on top of LLMs, and research all need different things, and beginner advice for one is wasted time for another.
What tends to work is picking one small thing to ship and learning only what it forces you to learn. A small app that answers questions over your own notes teaches prompting, embeddings, retrieval and evaluation in a couple of weeks. Fine-tuning a small model on the Hub with TRL teaches datasets, tokenizers and GPU limits.
Either way, let AI write the boilerplate and glue code. What you need to understand yourself is why an output is wrong: reading the data, checking the eval, noticing when the model is guessing.
What’s the first thing you’d want to build or be able to do once you got good at this?