ElixirConf 2023 - Sean Moriarity - MLOps in Elixir: Simplifying traditional MLOps with Elixir

Simplify traditional machine learning operations (MLOps) with Elixir, a modern language that simplifies deployment and management of machine learning models.

Key takeaways
  • Elixir simplifies traditional MLOps with its named entity recognition (NER) tool.
  • Named entity recognition is a text natural language processing task that replaces some $100 enterprise labeling tools.
  • With Elixir, you can implement pre-processing, inference, and post-processing for your machine learning models.
  • NX Serving is a serving abstraction that implements dynamic batching, encapsulating preprocessing, inference, and post-processing for you.
  • With NX Serving, you can integrate your machine learning models with your application without requiring expertise in machine learning.
  • Machine learning operations (MLOps) is the process of deploying machine learning models to production; Elixir simplifies this process with its serving abstraction.
  • Elixir has abstractions for machine learning, such as serving, that make it easier to implement and deploy machine learning models.
  • Livebook is a tool for machine learning in Elixir, allowing you to create, train, and deploy machine learning models in a notebook-like environment.
  • Ortex is a storage-only backend for machine learning models, allowing you to store and retrieve your models without having to worry about the underlying infrastructure.
  • Machine learning models can be wrapped in an NX Serving, which takes care of the concurrency for you and allows you to use the resources you have more efficiently.
  • With Elixir, you don’t need to be a machine learning expert to deploy machine learning models in production.
  • Elixir can be used to implement labeling tools, which can replace some $100 enterprise labeling tools.
  • The ecosystem has a lot of room for growth, and the author is excited to see where things will go from here.
  • The author’s book “Genetic Algorithms in Elixir” was a spark in the community’s interest in machine learning.
  • LiveView can be used to connect to your application and display real-time updates.
  • The author is a huge sports fan and can be found on Twitter (@Sean_Moriarty).
  • The author also has a Twitter-like application called X, which showcases his work with Elixir.