ML Compiler & System
Core
Building a high-performance open-source Python compiler (Quadrants) that translates Python code into optimized machine code for CPU and GPU backends, specifically targeting physics and performance-critical numerical computing.
Role type
Senior IC ML Compiler & System Engineer
Builds
Quadrants compiler infrastructure, including API, IR, lowering, optimization passes, and numerical primitives for CPU (x86, Arm64) and GPU (CUDA, AMDGPU, Metal, Vulkan) backends.
Domain
High-performance computing, compiler technology, GPU architecture, numerical computing
Deliverable
production ML models | product features
Required skills
GPU architecture (registers, shared memory, barriers, atomics, threading), compiler design (IR, lowering, optimization), Python API design, numerical primitives (linear algebra, solvers), system-level performance tuning
Preferred skills
CPU architecture intuition, experience with differentiable simulation, open-source project leadership
Technologies
Python, CUDA, PTX, C, inline assembly, NumPy, PyTorch
Responsibilities
Design and own the full compiler stack from API to backends, implement optimization passes for vectorization and memory efficiency, develop and benchmark numerical primitives, collaborate with the Genesis physics simulator team, drive open-source adoption and roadmap
Seniority
Senior, hands-on IC with potential to lead a small team
