I wrote a blog post on going from not knowing anything about deep learning last year to training state of the art OSS models - .
Hope it helps you.
tldr; read the deep learning book, implemented papers + taught, built open source tools
@VikParuchuri
would be interesting to read a more meta version of this post too. there's nothing stopping anyone else from doing what you did, why are you the only one who did it?
@CopetimusPrime
This is an interesting question. I would also guess, at the minimum, hundreds of people have taken a similar path (software engineer without deep learning experience, learn, get a research job). I guess the interesting part is the middle, though
@VikParuchuri
@VikParuchuri
the post is amazing and inspirational.
Can you share a bit about the effort (hrs/week?) you put in to follow such a cool journey?
For the rest of us, I'd probably 3X it, as a rule of thumb
@VikParuchuri
@VikParuchuri
- Great post! Two questions:
1. While doing LLM development, do you use any abstractions like DSPy, etc? Would you recommend any?
2. Hoping to read the section on "cleaning text data" in .
@ThereBeLyte
I haven't used DSPy, but wouldn't recommend any abstractions until you understand how things work under the hood. Prompt generators are really hard to debug.
@abhagsain
A decent amount, but mostly stats and linear algebra, with some calculus - it was easier to learn than I thought it would be (fear of math blocked me from studying deep learning for a long time)
@VikParuchuri
Inspirational post. Thanks for sharing and congrats on the new role! Your post really makes me think I should follow my passion, take a one year sabbatical to learn all I can, and contribute to open source projects I find interesting and valuable.