Principal Software Engineer, SDK & Lowering Stack

About the role

This position centers on constructing the essential software infrastructure for a novel dataflow architecture designed to accelerate generative AI workloads. You will join a specialized team responsible for the software development kit and the lowering stack that bridges high-level models to the underlying hardware. The work involves creating the primary interfaces and tooling that kernel authors and compiler engineers rely on daily. It is a foundational engineering role where your output directly dictates the usability and performance ceiling of the platform. You will operate at the intersection of compiler technology, runtime systems, and bare-metal hardware enablement.

Key facts

What you'll do

• Architect and implement the kernel authoring software development kit tailored for the Corsair dataflow architecture, ensuring it exposes hardware capabilities efficiently to developers.
• Establish and enforce rigorous software engineering standards across the codebase, including the design of continuous integration and continuous deployment pipelines, comprehensive automated testing frameworks, and mandatory code review workflows.
• Design intuitive, domain-specific application programming interfaces that serve the distinct needs of kernel developers and compiler backend teams, balancing abstraction with performance transparency.
• Build and maintain a suite of development utilities that streamline the kernel authoring lifecycle and facilitate the modeling lowering flows from graph representations to executable binaries.
• Provide technical mentorship and guidance to junior engineering staff, fostering a culture of technical excellence, knowledge sharing, and collaborative problem-solving.
• Perform deep analysis of the interplay between hardware microarchitecture and software stack behavior to identify bottlenecks and propose cross-layer optimizations.
• Own the end-to-end delivery lifecycle for critical software components, spanning initial requirements analysis and architectural design through implementation, integration, debugging, and production deployment.
• Drive the scaling of software deliverables to meet aggressive project timelines, coordinating cross-functional dependencies and managing technical risk throughout the development process.
• Collaborate closely with hardware architects to define and refine the instruction set architecture and programming model interfaces for current and future silicon generations.
• Develop and optimize the compiler lowering passes that translate high-level intermediate representations into optimized dataflow executables for the target architecture.
• Implement robust debugging and profiling tooling that provides visibility into kernel execution behavior and performance characteristics on the accelerator.
• Contribute to the technical roadmap and strategic planning for the SDK and lowering stack, aligning engineering investments with product milestones and customer requirements.

Requirements

• Advanced degree (Master's) in Computer Engineering, Mathematics, Physics, or a closely related technical discipline accompanied by a minimum of eight years of relevant industry experience, or a Doctorate with at least four years of experience.
• Expert-level proficiency in modern C++ and Python within a Linux development environment, demonstrating fluency with standard toolchains, debuggers, and build systems such as CMake or Bazel.
• Demonstrated track record of instituting and maintaining software engineering best practices, specifically the implementation of CI/CD infrastructure, continuous testing methodologies, and rigorous peer review processes in production environments.
• Deep foundational knowledge of computer architecture principles, advanced data structures, operating system internals, and the mathematical fundamentals underpinning machine learning algorithms.
• Proven experience designing and shipping high-quality, scalable SDKs and APIs from a systems engineering perspective, with attention to versioning, backward compatibility, and developer ergonomics.
• Self-directed work style with a strong sense of ownership, capable of leading complex technical initiatives autonomously while collaborating effectively across multidisciplinary teams.
• Demonstrated ability to debug complex, multi-layered software stacks spanning user-space applications, kernel drivers, and firmware interactions.
• Experience with build system architecture and dependency management for large-scale, multi-repository codebases.

Nice to have

• Direct experience authoring and optimizing high-performance compute kernels for graphics processing units, tensor processing units, or custom machine learning accelerators using assembly or intrinsic programming.
• Background in High-Performance Computing environments, including workload scheduling, distributed memory management, and resource optimization across large clusters.
• Hands-on development experience with machine learning compiler infrastructures, intermediate representations, or optimization algorithms, specifically within frameworks such as MLIR, LLVM, TVM, or Glow.
• Familiarity with leveraging AI-assisted coding tools and large language models to accelerate development velocity and enforce code quality standards.
• Prior tenure in a startup, incubation lab, or small high-autonomy team environment where rapid iteration and broad technical scope were required.
• Professional experience at a major cloud provider or within an AI compute and subsystem organization, providing insight into datacenter-scale deployment constraints.

Skills & tools

• C/C++
• Python
• Linux
• CI/CD
• MLIR
• LLVM
• TVM
• Glow
• CMake
• Bazel
• Git
• gdb
• perf
• Docker
• Kubernetes

Practical notes

This is a hybrid role anchored in Santa Clara, California, though the organization offers flexibility for candidates located elsewhere in the United States or Canada. The company maintains a policy of not accepting unsolicited resumes or candidate profiles from external recruiting agencies; all applications must be submitted directly through the official channels. Visa sponsorship details are not explicitly stated in the source material; candidates requiring work authorization should verify eligibility directly with the hiring team during the application process. The compensation package includes a base salary range of $190,000 to $300,000 supplemented by equity grants and performance-based bonuses.