About SLM
SLM is a compute infrastructure company building supercapability systems. We're developing new classes of high-performance architectures that deliver advanced performance, precision, and physical resilience in extreme environments. Many of our operating conditions are where conventional computing fails: radiation fields that destroy standard electronics, temperature extremes, and power constraints that demand every design decision be justified.
Our first-generation systems target scientific instrumentation with hard deployment windows that won't reopen for years or close to more than a decade. The engineering challenges are significant: advanced semiconductor processes pushed to their limits, specialized memory technologies integrated in ways they haven't been before, multi-die systems coordinated through custom packaging, and software toolchains that extract maximum performance from novel hardware. We're a small team solving problems most engineers never get close to, move at a high cadence and with rigor, and ship systems that need to work the first time in environments where you can't iterate after deployment.
About this role
You'll build production compiler infrastructure for spatial computing architectures with deterministic real-time guarantees. This means creating the software toolchain and programming model that allows application developers to achieve high performance without writing hardware-specific code.
This involves designing the compilation strategy from high-level languages through intermediate representations to configuration bitstreams that program the hardware. You'll implement memory-hierarchy-aware optimization passes that exploit locality across cache levels and specialized non-volatile memory, develop static scheduling algorithms that provide bounded latency guarantees, and create dataflow analysis frameworks that map computations to spatial processing element arrays.
The technical challenges go beyond standard compiler development. You're compiling for a spatial architecture where computation maps to physical locations, where data movement has explicit cost that dominates computation time, where memory hierarchy has multiple levels with very different latency and bandwidth characteristics. You'll implement polyhedral optimization for loop transformations, develop custom intermediate representations that capture spatial dataflow, and create backend code generation that maps to hardware configuration rather than instruction sequences.
You'll work closely with the architecture team to ensure the compiler targets the hardware being built, with the algorithms team on expressing computational patterns the compiler can recognize and optimize, with application developers on language features and performance debugging tools they need. You'll make fundamental design decisions about compilation strategy that affect both what applications can express and what performance they can achieve.
The timeline for this system is demanding. The compiler needs to be functional before silicon arrives so application developers can prepare code, mature enough for performance evaluation when first systems deploy, and production-ready for customer use. You'll make pragmatic decisions about which optimizations to implement first, which language and reliability features are essential versus nice-to-have, and where to focus optimization effort for maximum impact.
What we're looking for
- Advanced degree in Computer Science or equivalent depth from building production compiler systems.
- You've developed significant compiler infrastructure — not just used LLVM but implemented optimization passes, designed intermediate representations, or built backend code generators.
- You understand dataflow analysis, optimization algorithms, and the tradeoffs between compilation time and generated code quality.
Experience with spatial compilation, dataflow languages, or CGRA toolchains is particularly relevant. If you've worked on compilers for accelerators, GPUs, or other specialized hardware, you understand the challenges of compiling for non-von-Neumann architectures. Familiarity with real-time systems or safety-critical software helps contextualize the determinism requirements. Strong programming skills in C++ and comfort with large software systems are essential.
We're looking for someone who combines deep compiler expertise with pragmatic engineering judgment, who can design elegant abstractions but also ship working code under schedule pressure, and who communicates effectively about technical tradeoffs and schedule risks.
What we offer
As an early team member, you'll shape capabilities and systems with first-order consequences for the future and direction of humanity's enterprise.
This is accompanied by a strong equity package, competitive base salary, and comprehensive benefits including enhanced healthcare coverage for you and your family, robust family planning support, life insurance, flexible time off and paid holidays, retirement plans with matching, daily meals at our headquarters, and relocation support.
Our primary operational base is set in the Bay Area, and our labs are headquartered in a part of the city set beside cypress groves and coastal trails. Think natural light, fresh ocean air, and panoramic views. We work intensely but deliberately invest in removing avoidable frictions from your life so you can dedicate maximum bandwidth to your core work.
If we make you an offer, we will work hard to get you onto our team and can even sponsor visas and green cards once eligible.
We strongly encourage you to apply even if you feel you don't meet every qualification or attribute as described. We care more about evidence of strong ability and a high signal-to-noise ratio.
Role details
- Category: Compiler & Software
- Role: Compiler Architect / Compiler Engineer
- Work type: On-site
- Employment: Full-time
- Location: Bay Area, California
- Salary range: $220,000 - $300,000