Software Engineer - Compilers

Software Engineer - Compilers

We’re working with a well-funded Series A company rethinking how AI workloads actually run.

Today, AI systems are tightly coupled to specific hardware, which creates massive inefficiencies in cost, scaling, and performance. 

This team is breaking that model. They’re building a system that decomposes workloads and executes them across heterogeneous compute (GPUs, accelerators, multi-gen hardware), automatically optimising for performance and efficiency.

What you’ll do?

• Own core execution systems from design → build → deployment → operation
• Work on compilers, runtimes, and execution layers for large-scale AI workloads
• Translate high-level workloads into efficient, hardware-aware execution
• Optimise performance across diverse and evolving hardware environments
• Make hard tradeoffs across latency, throughput, and cost
• Debug and improve behaviour in complex, production systems

What makes this interesting?

• You’re working on the critical path of AI execution — not tooling around it
• Direct impact on performance and cost (this is not abstract optimisation)
• Deep technical problems across compilers, runtimes, and hardware interaction
• Heterogeneous compute: designing systems that go beyond “just use GPUs”
• Early-stage: real ownership and the ability to shape fundamental architecture
• Serious engineering environment: decisions are debated, not hand-waved

We’re looking for:

• Engineers with strong systems fundamentals (compilers, runtimes, or low-level infra)
• Experience working on performance-critical or execution-layer systems
• Evidence of owning and shipping complex technical work
• Ability to reason clearly about tradeoffs and system behaviour
• Comfort operating in ambiguous, evolving problem spaces