Modular is on a mission to revolutionize AI infrastructure by rebuilding the AI software stack. As a Mojo Libraries Engineer, you will design and implement core abstractions for the Mojo programming language, enabling developers to build efficiently and optimize AI systems.
Responsibilities:
- Partner with teams across Modular—including compiler, kernel, and runtime engineers—to design APIs that are ergonomic, expressive, and performance-conscious. Your work will directly shape the developer experience and deliver high business impact by enabling critical customer workloads and accelerating platform adoption
- Design and implement abstractions that enable the development and optimization of kernels and algorithms running on GPUs and CPUs, delivering top-tier performance and accuracy throughout the GenAI inference solution stack
- Optimize library functionality for performance, memory efficiency, and hardware utilization
- Participate in design discussions and code reviews to uphold high engineering standards
- Collaborate with other engineering teams to align the standard library with broader platform needs
- Contribute to a fully open source project — everything you build will be public and part of our GitHub repo
- Engage with the open source community to support high-quality contributions and foster a welcoming, inclusive environment
Requirements:
- Experience with GPU programming languages like CUDA or OpenCL
- Deep understanding of GPU architecture (memory hierarchies, tensor cores, etc.) and how it influences algorithm and API design
- Proficiency in modern systems languages (C++, Rust, or Swift), with strong skills in low-level programming, performance optimization, and memory management
- Passion for both library design and performance, with a commitment to following modern API patterns and language evolution in C++, Rust, and Swift ecosystems
- Hands-on open source experience — either through meaningful contributions or by maintaining and reviewing community submissions
- A collaborative mindset, intellectual curiosity, and drive to solve complex technical challenges as part of a high-performing team
- Strong alignment with Modular's cultural values and enthusiasm for building a thoughtful, inclusive engineering culture
- Familiarity with GPU assembly languages such as PTX and SASS
- Knowledge of cutting-edge APIs such as CUTLASS or Triton
- Prior experience with Mojo, especially with accepted contributions to the Mojo Standard Library
- Familiarity with LLVM, MLIR, or Python, particularly in systems or compiler-related contexts
- An advanced degree in Computer Science or a related field