Compiler Backend Engineer for AI Inference Hardware

Normal Computing Corporation; Mountain View (CA); On-site; USD 180,000 - 280,000; Full time

Job summary

Normal Computing in Mountain View is seeking a Compiler Engineer focused on the backend to map LLM inference onto our custom hardware. You will work at the boundary between the compiler and hardware, designing and evaluating compilation strategies against simulators and FPGA prototypes.

You will own code generation, scheduling, fusion, and hardware co-design, collaborating with frontend engineers and hardware teams to ship measurable performance improvements.

Qualifications

• Experience with compiler optimization for accelerators or non-CPU targets.
• Track record of shipping optimizations that improve performance.
• Comfort designing compilation approaches from first principles.

Responsibilities

• Code generation: build passes to convert lowered programs into hardware-executable code.
• Scheduling and memory planning: optimize data and work placement for low latency.
• Fusion and dispatch: decide on fuse/dispatch to maximize compute.
• Hardware co-design: feed compiler constraints back to hardware architects.
• Profiling and validation: verify improvements with simulators, FPGA, and silicon.
• Cross-team handoffs: determine operations for compiler vs hand-tuned kernels.

Skills

Compiler optimization

Code generation

Scheduling

Fusion

Resource allocation

Hardware awareness

Performance measurement

Tools

Python

C

C++

Rust

Job description

Normal Computing in Mountain View is seeking a Compiler Engineer focused on the backend to map LLM inference onto our custom hardware. You will work at the boundary between the compiler and hardware, designing and evaluating compilation strategies against simulators and FPGA prototypes.

You will own code generation, scheduling, fusion, and hardware co-design, collaborating with frontend engineers and hardware teams to ship measurable performance improvements.