Runtime Engineer (C/C++)

Evollabs Tech; Dubai; On-site; AED 400,000 - 650,000; Full time

Job summary

Evollabs Tech is seeking an experienced Senior Runtime Software Engineer to own the core runtime architecture that bridges compiler frameworks and AI hardware. You will drive runtime performance and stability, and shape memory management, scheduling, and observability—collaborating with firmware and driver teams.

You will work on concurrency models for model execution, profiling, and low-latency optimizations to ensure scalable, maintainable solutions for real-world AI workloads on our NPUs.

Qualifications

• 5–8+ years in systems software, runtime or performance-critical infra.
• Proficiency in C/C++ and Python.
• Strong knowledge of OS fundamentals, multithreading, synchronization and memory management.
• Experience near hardware (GPU/accelerator/driver/firmware).
• Familiarity with command queues, execution engines and DMA-based systems.
• Understanding of computational graphs, tensor execution and memory layouts.
• Experience profiling/optimizing latency and throughput in performance-sensitive systems.

Responsibilities

• Own core runtime architecture bridging compiler and hardware.
• Improve runtime execution engine performance and stability.
• Design device memory management and efficiency improvements.
• Collaborate on command submission and hardware interaction layer.

Skills

C/C++

Python

Operating systems

Multi-threading

Memory management

Profiling/Optimization

GPU/driver/firmware familiarity

Tools

CUDA

MLIR

LLVM

ONNX Runtime

TensorRT

Job description

About Us

We are a tech company specializing in the design and development of cutting-edge, customized server hardware solutions optimized for artificial intelligence and machine learning applications. Our mission is to empower businesses and researchers to accelerate their AI initiatives by providing them with high-performance, scalable, and energy-efficient hardware infrastructure.

As a rapidly growing company at the forefront of AI hardware innovation, we are constantly seeking talented and motivated individuals to join our team. We offer a dynamic and challenging work environment, with opportunities to make a significant impact on the future of AI technology.

You’ll Collaborate With

Compiler engineers, firmware & driver engineers, hardware architects, ML framework engineers and performance & validation teams

What You’ll Own

• The core runtime architecture that bridges the compiler and the hardware
• Runtime execution engine performance and stability
• Device memory management design and efficiency
• Command submission and hardware interaction layer (in collaboration with firmware/driver teams)
• Concurrency and scheduling model for model execution
• Observability tooling (profiling, logging, tracing) within the runtime
• Technical design decisions that ensure long-term scalability and maintainability

Minimum Qualifications

• 5–8+ years of experience in systems software, runtime systems, or performance-critical infrastructure
• Strong proficiency in C/C++and Python
• Solid understanding of fundamentals of operating systems, multi-threading, synchronization, memory management, and cache behavior
• Experience working close to hardware (GPU, accelerator, driver, or firmware environments)
• Familiarity with command queues, execution engines, and DMA-based systems
• Understanding of computational graphs, tensor execution, and memory layouts
• Experience profiling and optimizing latency and throughput in performance-sensitive systems

Preferred Qualifications

• Experience building or architecting a runtime from scratch
• Familiarity with MLIR, LLVM, or compiler-runtime interfaces
• Experience with one or many from CUDA, ROCm, TensorRT, TVM, XLA, ONNX Runtime or similar runtimes
• Experience integrating custom hardware backends into PyTorch, ONNX or OpenXLA
• Knowledge of quantization and mixed-precision inference
• Experience with multi-device or distributed execution
• Background in AI accelerator or semiconductor environments

What Success Looks Like

• You enable reliable end-to-end execution of compiled models on our NPU
• Runtime overhead is minimal and performance targets are consistently met
• Memory management and scheduling are efficient and scalable
• The runtime architecture is robust, maintainable, and extensible
• Compiler, firmware, and hardware teams can build confidently on top of your execution layer
• The system supports real-world AI workloads with stability and observability

Join us in our mission to democratize AI compute — where your firmware expertise becomes the bedrock of tomorrow's AI breakthroughs.