Given the meteoric rise of ML/AI in the past few years, I'm surprised that Kaggle doesn't come up more often. It was all the rage 2013-2018...then most I heard about it was that it allowed free access to TPUs.
I don't see that as a universally bad thing. I feel this is actually very representative of the majority of ML projects that most organizations encounter. Most organizations don't have petabytes of data and a huge compute budget to train a DNN. They typically have megabytes to gigabytes of somewhat crappy data and need something that can be developed and deployed relatively quickly for low cost.
Platforms like Kaggle, DrivenData, Zindi, AIcrowd, CodaLab and others are running dozens-hundreds of competitions a year in total, including ones linked to top academic conferences. One interesting recent one is this one on LLM efficiency - trying to see to what extent people can fine-tune an LLM with just 1GPU and 24h: https://llm-efficiency-challenge.github.io/
Or the Makridakis series of challenges, running since the 80s, which are a great testbed for time-series models (the 6th one finished just last year): https://mofc.unic.ac.cy/the-m6-competition/