BS in Computer Science or Electrical Engineering with 15+ years experience, or MS with 12+, or PhD with 10+ years.
12+ years in front-end compiler and systems software development focused on ML inference.
Deep experience with compiler technologies such as MLIR, LLVM, or Torch-MLIR.
Strong understanding of model optimization techniques for inference including quantization and scheduling.
Expertise in deploying ML models to heterogeneous cloud environments.
Experience with AI frameworks like PyTorch or TensorFlow.
Responsibilities
Architect the MLIR-based compiler for cloud inference workloads.
Lead development of compiler passes for model partitioning and optimization.
Design support for hybrid compilation flows with adaptive resource utilization.
Define compiler abstractions for efficient cloud deployment.
Drive performance through IR transformations and distributed execution strategies.
Mentor a team of compiler engineers to optimize inference software.
Benefits
Hybrid work model allowing onsite work 3-5 days per week or fully remote.
Inclusive culture promoting respect and collaboration.
Opportunity to shape innovative AI technologies.
Leadership development and mentoring opportunities.
Focus on work-life balance in a progressive work environment.
Full Job Description
Location:
Santa Clara, CA, headquarters, but open to other locations within the US/Canada.
The role: Principal Software Engineer - SDK & Lowering Stack
What you will do:
Help architect and develop the kernel authoring SDK that targets our novel Corsair dataflow architecture.
Champion software engineering best practices across the team by establishing robust CI/CD pipelines, driving automated testing, and enforcing rigorous code review processes.
Build intuitive, domain-specific APIs, determining the optimal boundaries for kernel developers and compiler teams.
Create development tools to be used in kernel authoring and modeling lowering flows.
Lead and mentor junior engineers, fostering a culture of technical excellence, accountability, and collaborative problem-solving.
Leverage systems engineering ability to understand complex interactions between hardware and software components, proposing and implementing solutions integral to the software stack.
Take ownership of delivering solutions by interacting with stakeholders, generating requirements, documenting designs, implementing, debugging, and deploying.
Build and scale software deliverables in a tight development window.
Minimum experience:
MS in Computer Engineering, Math, Physics, or a related degree with 8 years of industry experience or a PhD in Computer Engineering, Math, Physics, or a related degree with 4 years of industry experience.
Proficient in C/C++ and Python development in a Linux environment and using standard development tools.
Proven track record of driving software engineering best practices, including CI/CD integration, continuous testing, and leading comprehensive code reviews.
Deep grasp of computer architecture, data structures, system software, and machine learning fundamentals.
Demonstrated experience designing high-quality, scalable APIs and SDKs, with a comprehensive systems engineering mindset to solve end-to-end problems.
Self-motivated team player with a profound sense of ownership and leadership.
Preferred experience:
Proven experience writing and optimizing high-performance kernels targeting GPUs, TPUs, or custom ML accelerators.
Background working in High-Performance Computing (HPC) environments, managing large-scale distributed workloads and resource optimization.
Experience developing ML compilers and working with intermediate representations or algorithms (e.g., MLIR, LLVM, TVM, Glow, etc.).
Familiarity with leveraging AI-assisted coding tools to accelerate software development and elevate code quality.
Prior startup, small team, or incubation experience.
Work experience at a cloud provider or AI compute/subsystem company.