Runtime Lead
Efficient Computer
·Today
- Salary
- $220k – $270k/yr
- Location
- San Francisco, Bay Area OR Pittsburgh, PA · San Francisco, California, United States
- Department
- System Software
- Seniority
- Lead
- Experience
- 10+ years
- Education
- Master
- Source
- Greenhouse
Description
Efficient Computer is rethinking computing from the ground up to solve one of AI's biggest challenges: energy. Our Fabric architecture delivers a 10–100x improvement in energy efficiency for general-purpose computation, including AI, while supporting familiar programming languages and software frameworks—unlocking new possibilities for autonomous machines, intelligent infrastructure, wearables, and space systems. With our first processor, Electron E1, now shipping in volume and more than $97 million in Series B financing recently announced, we're accelerating customer deployments and scaling our architecture to data center performance. Join us to build the hardware and software that bring intelligence wherever it's needed and help shape the next generation of computing.
We are looking for a Runtime Lead to architect and design the runtime software stack that brings the Fabric to life, from the Linux kernel-mode driver through the user-mode driver and runtime library that sit between the effcc compiler and the Fabric ABI. You will set the overall technical direction, plan the technical work with engineering management, collaborate with partner teams, and provide technical leadership to the engineers who build it, while staying hands-on in the code. You will take the stack from pre-silicon environments, involving QEMU, functional simulation, emulation, and FPGAs, through first silicon to ensure applications run on day one.
This is a key role for the System Software team. As such, you will represent the runtime team to the compiler, architecture, digital design, verification, and product teams during all phases of development.
Key Responsibilities
Provide Technical Leadership for the Runtime Team
- Provide technical leadership to the engineers who own the kernel-mode driver and the user-mode runtime: guide technical priorities and ownership with the engineering manager, mentor engineers, and help interview as the team scales.
- Own the runtime technical roadmap: scope, milestones, and estimates. Work with the engineering management on the tradeoffs needed to hit committed dates, and raise technical risks early.
- Set the engineering bar for the team: design reviews, code reviews, coding standards, test strategy, and release practices.
- Build consensus on technical decisions within the team and with partner teams.
- Stay hands-on, contributing code and design to the most critical parts of the stack.
Architect the Runtime Software Stack
- Own the end-to-end architecture of the runtime software stack: the kernel-mode driver, the HAL and Fabric ABI, the user-mode driver and runtime library, and the interfaces used by ML framework backends (e.g., PyTorch).
- Help define the split between kernel and user space: what belongs in the driver, what belongs in the runtime,and how the two share memory, synchronization primitives, and error reporting.
- Help define the unified memory model between ARM cores and the Fabric accelerator, including SMMU (IOMMU) integration, tensor pool allocation, shared virtual addressing, and zero-copy data paths.
- Help define, with the compiler team, the in-memory representations generated by the compiler and consumed by the Fabric, along with binary formats, program loading, and the launch ABI.
- Help define, with the help of the firmware and architecture teams: the multi-island task scheduling policy, microcode loading, and the software stack for the embedded risc core.
- Keep the stack layered and portable across device generations, and across QEMU, functional simulation, emulation, FPGA, and silicon backends.
Represent the Team Cross-Functionally
- Represent the runtime team in architecture, digital design, compiler, embedded, verification, and product forums, and bring the software point of view into hardware decisions before tapeout.
- Keep engineering leadership and program management informed on technical status, risks, and decisions.
- Work with product and application teams to turn SDK, framework, and application requirements into runtime priorities.
Drive Pre-Silicon Bringup and Platform Enablement
- Plan and lead bring-up across the full stack, from first boot in QEMU and functional simulation through emulation, FPGA prototypes, and first silicon, coordinating kernel and user-mode work.
- Own the runtime quality strategy, including driver and API unit tests, conformance tests, emulation-based regression suites, performance benchmarks, and hardware-in-the-loop validation.
- Serve as the escalation point for the hardest debug problems across kernel and user space.
Required Qualifications
- B.S. in Electrical/Computer Engineering, Computer Science, or equivalent; M.S. or PhD preferred.
- 10+ years of systems software experience, including significant hands-on Linux kernel-mode driver and user-mode runtime library development for accelerators (GPUs, NPUs, DSPs, FPGAs, or similar).
- Experience as a technical lead: planning and estimating technical work, mentoring engineers, and delivering against hardware and silicon schedules.
- Experience architecting significant components of a production hw accelerator software stack, including the interfaces between kernel and user space and the public runtime APIs: ABI and HAL design, versioning, error models, device telemetry, thread safety, and synchronization.
- Deep understanding of memory management across kernel and user space: cache coherence, IOMMU/SMMU configuration, unified or shared virtual memory, allocators, and zero-copy data paths.
- Experience working with compiler teams on binary formats, loaders, JIT vs AOT compilation, launch ABIs, and kernel metadata.
- Strong written and verbal communication; able to represent the team in cross-functional settings and executive reviews, and to influence decisions across hardware and software teams.
- Expert C and C++ and strong Python skills; comfortable working with AI-assisted development tools for code, analysis, and debugging.
Desired Qualifications
- Experience contributing to or leading development of a runtime stack, e.g., CUDA (cudart), ROCm (HIP, ROCr, HSA), Tenstorrent runtime (TT-Metalium), oneAPI Level Zero, OpenCL, or similar accelerator runtime software stack.
- Experience as a technical lead on a first-generation silicon product, where the software stack and the hardware mature together.
- Familiarity with LLVM or MLIR, and with how compilers for accelerators hand off to runtimes.
- Knowledge of dataflow, spatial, or other novel compute architectures, and of how compilers and runtimes target them.
- Experience with FPGA prototyping and hardware emulation platforms (Palladium, Zebu, Veloce, or similar).
- Experience with pre-silicon and first-silicon bring-up on ARM platforms: QEMU, emulation, or simulation environments; memory-mapped I/O, interrupt handling, and the Linux boot flow; and the ARM architecture.
We offer a competitive base salary for this role, generally ranging from $220,000 to $270,000, plus a 10% annual bonus, meaningful equity, and comprehensive benefits. Final compensation will depend on experience, level, and location, with flexibility for the right candidate.
Why Join Efficient?
Build what's next. At Efficient, you'll tackle big challenges with room to take ownership and make an impact. We back you with competitive pay, equity, a 401(k) match, company-paid benefits, paid parental leave, and flexibility—plus opportunities to grow your skills and career as we scale.