Getting Started with ML
I've found a few different resources and have decided to spend the next few months diving deep into ML while I look for a job. Integrating with AI products is cool, but it's much cooler to know how these things work under the hood.
For the most part, this process mirrors other learning I've done in the past. Typically that consists of scouring the internet for content, consuming as much of it as possible, and then building something to see where it shakes out. I've done some form of this for my entire career and the truth is that it can be very easy to over-engineer your path forward.
So I'm starting with a few simple resources and plan to build things as I go!
- FastAI's free course + accompanying FastBook
- Karpathy's GPT Zero to Hero
- Kaggle Learn
I'll be documenting everything that I learn as I continue to take notes in Obsidian, and while I'm at it I'll build a workflow to use Jupyter Notebooks.
Upcoming known projects I'll be trying to take on as time allows, besides just going through the course:
- AI-generated calendar briefings
- gitRAG - a way to index + use natural language to query your own git history
- Overcommunicator - automate the job hunt with tracking + email sequencing
- MinimapReader - watch the League of Legends minimap footage to detect ganks & more!
I'm hoping by the end of this a few things will be a lot less mystifying - diving deeper into computer vision, agentic software, RAG, training my own models, and more!