Mojo: Programming language – Combine Python and MLIR for AI development

Recently, Modular AI – a new company founded by Chris Lattner, co-founder of LLVM and Swift programming language – released a new programming language called Mojo.

What is Mojo?

Mojo is built on top of Python, specifically designed for Artificial Intelligence applications, and as showcased multiple times, it is 35000 times faster than Python. Data scientists, including Jeremy Howard – the founding research fellow of – have acclaimed this to be “the biggest progress in the programming language in decades”.

A quick look at the impressive background of its founder – Chris Lattner

Before we jump into what Mojo is all about, let us take a quick look at the impressive background of its founder – Chris Lattner:

  • One of the major project initiators and authors of the LLVM Project
  • Author of the Clang Compiler
  • Creator of Apple’s Swift Programming Language; worked 11 years at Apple leading their Developer Tools Department
  • Worked with Tesla’s Autopilot Team and Google’s TensorFlow Group.

In 2022, Chris Lattner co-founded the developer-first AI platform Modular AI where he also serves as its CEO. Mojo language is Modular AI’s newest release which is again tailor-made for AI developers. Combining the ease of usage from Python with a powerful performance from C language, it has massive advantages when it comes to programmability on AI hardware and the scaleability of AI models.

People Also Read: AI Innovation with Microsoft Bing and Edge

The main features of Mojo include:

  1. Python Like Syntax: Similar syntax to Python making it easy for developers to pick up easily
  2. Compatibility with Python Libraries: compatible with many existing Python scientific computing libraries
  3. Superior Performance: 35000x Faster than Python
  4. Single Language Writing: need not master C++ or Python separately; carry out both functionalities using just Mojo
  5. Parallel Processing: MLIR used by Mojo for vectorization, threading, and processing on dedicated AI hardware units
  6. Easier Extension Of Models: makes upgrading models easier than ever before

Currently still in the development process yet fully available to try out in the JupyterHub test environment. Apply now here


We will be happy to hear your thoughts

Leave a reply

Hug Techs