Why does Python get to have all the AI and ML fun?

What if there was a better way to build the applications that serve these models? A way that offered unparalleled concurrency, fault tolerance, and low-latency performance out of the box?

Ever thought to yourself, the world needs a searchable Dota 2 audio clips app, so you can also troll my colleague, Maicon with meme-able voice clips from one video game in particular?

No? Ok, maybe its just a me thing.

In this session, we'll introduce and explore how the Elixir language is almost purpose built to troll your work mates with AI applications. You'll see how we use Elixir's Bumblebee library to plug in powerful AI models and how the NX library makes it all possible.

Next, we’ll leverage Elixir's legendary foundation for stability (called OTP) to effortlessly build a robust data processing pipeline. You'll leave understanding not just how to build a smart search application, but why Elixir's unique features make it a rising star for deploying dependable AI in the real world.

Even if all of this sounds like gibberish to you, come along and see how we can build the dumbest usage of linear algebra imaginable.