Hi everyone, I’m the creator of Made With ML and I wanted to share that V1 of the open-source course is finally complete! We cover topics across data → modeling → serving → testing → reproducibility → monitoring → data engineering + more, all with the goal of teaching how to responsibly develop, deploy and maintain production ML applications.
- 🛠 Project-based
- 💡 Intuition (first principles)
- 💻 Implementation (code)
- 🏆 30K+ GitHub ⭐️
- ❤️ 40K+ community
- ✅ 49 lessons, 100% open-source
[Background] I started Made With ML as a way for me to share my learnings from the different contexts I’ve brought ML to production in the past. I currently work closely with teams from early-stage/F500 companies, as well as collaborating with the best tooling/platform companies, to make delivering value with ML even easier and faster.
[Request] I keep all the lessons updated as I learn more (especially constantly evolving spaces such as testing and monitoring ML). But what are some modeling-agnostic topics that are missing here that are very crucial to production ML / MLOps? A few high priority ones on the TODO list include bias (identifying, mitigating), distributed workflows (not just for training), etc. What else should be added here?