Compiler Engineer (AI)

Location: London (Hybrid)

hirify.global is building high-performance infrastructure for numerical computing, including the open-source Quadrants compiler and the Genesis physics simulator.

What you will do

• Own the end-to-end development of the Quadrants compiler, including API design, IR, lowering, and optimization passes.
• Develop and maintain numerical primitives such as matmul, Cholesky, and solver kernels.
• Optimize naive Python code into high-performance kernels by implementing vectorization, register management, and cooperative threading.
• Collaborate with the Genesis physics simulator team to identify performance bottlenecks and drive API improvements.
• Benchmark real-world workloads to track performance and catch regressions early.
• Design and maintain the project roadmap while engaging with the open-source community.

Requirements

• Deep expertise in GPU architecture, including access patterns, registers, shared memory, barriers, and atomics.
• Strong proficiency in idiomatic Python and API design principles.
• Experience with compiler design, specifically the path from authoring to kernels (IR, lowering, optimization).
• Ability to reason about whole systems and manage technical complexity.
• Proactive, self-directed work style with a focus on ownership and technical debt reduction.
• Must be based in or able to work from London in a hybrid capacity.

Culture & Benefits

• Opportunity to define the state of the art in performance-critical numerical Python.
• Work on a high-impact, open-source project with a clear, ambitious roadmap.
• Environment that prioritizes usability and clean API design alongside raw performance.
• Focus on deep understanding of the codebase rather than just feature stacking.
• Support for using AI agents as productivity tools while maintaining full code ownership.