Made this free site that teaches everything required to get a job as either inference engineer or for model training by analyzing a bunch of current job requirements for job postings by top labs.
I'm working through it on my own but would love any advice or feedback.
It claims to start from "bedrock" but beings with a Horace He post that clearly assumes you've done this before. Its fine if you want to be an intermediate course, but you should begin by defining your audience, so that you can reach the correct audience. You should have a section explaining who this is for, and if its _not_ an intro course, you should say that (and probably avoid terms like "fundamentals" and "bedrock").
Or if this is an oversight and you intend this actually be a course that starts from zero, you shouldn't start with an article that's starts off talking about training loss, test loss, and overfitting, without defining those terms.
I think a curriculum like this is neat and might help with interviews since you go wide and have a checklist of things that you need to learn.
I'm on a totally different path for learning inference engineering. I self-host a voice AI app that has ~2000 monthly active users on my own GPU box.
This forces me to learn about production serving, KV cache, quantization, inference engine, observability and economics, prefill optimization since I'm optimizing for TTFT instead of decode speed, and many more.
It's fun since every optimization you do directly translate to a better user experience or allow you to serve more users using the same hardware.
Made this free site that teaches everything required to get a job as either inference engineer or for model training by analyzing a bunch of current job requirements for job postings by top labs.
I'm working through it on my own but would love any advice or feedback.
It claims to start from "bedrock" but beings with a Horace He post that clearly assumes you've done this before. Its fine if you want to be an intermediate course, but you should begin by defining your audience, so that you can reach the correct audience. You should have a section explaining who this is for, and if its _not_ an intro course, you should say that (and probably avoid terms like "fundamentals" and "bedrock").
Or if this is an oversight and you intend this actually be a course that starts from zero, you shouldn't start with an article that's starts off talking about training loss, test loss, and overfitting, without defining those terms.
I think a curriculum like this is neat and might help with interviews since you go wide and have a checklist of things that you need to learn.
I'm on a totally different path for learning inference engineering. I self-host a voice AI app that has ~2000 monthly active users on my own GPU box.
This forces me to learn about production serving, KV cache, quantization, inference engine, observability and economics, prefill optimization since I'm optimizing for TTFT instead of decode speed, and many more.
It's fun since every optimization you do directly translate to a better user experience or allow you to serve more users using the same hardware.