Modern ML engineers and platform developers know that scaling machine learning workloads isn’t a matter of just adding GPUs. Optimizing ML system performance requires orchestrating compute, storage, and data pipelines to maximal efficiency under production constraints. As models grow more complex, what once was processed within a single rack now requires distributed GPU clusters, shared … Read more
Rails 8 introduced support for multiple databases, and by default, it wants to create separate database instances for your primary data, cache, queue, and ActionCable connections. If you’re deploying to Heroku and you just follow the defaults, you’ll end up with four separate Postgres databases, which means four separate billable databases instances. I’m using a … Read more
So, I broke production today. Not in a small way either. I’m writing this because if you’re running Rails 8 with SolidQueue on Heroku, you might run into the same thing, and I’d rather you learn from my mistake than make it yourself. The Setup I’ve been working on a Rails 8 app that uses SolidQueue for … Read more
Every machine learning engineer eventually encounters this challenge: a model that performs perfectly in a notebook often fails in production. The problem isn’t the algorithm itself; it’s everything surrounding it. In a lab environment, data is clean, schemas are consistent, and dependencies remain stable. However, in a production environment, data changes daily, infrastructure evolves, and … Read more