Low Level Optimization Engineer

We are looking for exceptional engineers with deep low-level expertise to join our team

in optimizing and scaling next-generation AI systems.

This role sits at the intersection of machine learning, systems, and hardware, and

involves working on highly performance-critical components of our video generation

stack. You will play a key role in driving efficiency, throughput, and latency

improvements across the full software–hardware stack.

What You’ll Do

• Design and implement high-performance components for large-scale generative models

• Optimize compute-intensive workloads across CPUs, GPUs, and AI accelerators

• Develop custom kernels and low-level primitives for critical model paths

• Analyze and improve system-level performance (latency, throughput, memory utilization)

• Collaborate across ML, systems, and infrastructure teams to co-design efficient solutions

• Contribute to scaling real-time video generation pipelines to production environments

Requirements (Must Have)

• Strong passion for hardware–software co-optimization

• Proven experience in at least one of the following:
  Low-level programming (C/C++/CUDA/assembly or similar)
  Performance optimization and profiling of complex systems
  Hardware-aware software development
  SW-aware ASIC / architecture design

• Deep understanding of system architecture (memory hierarchy, parallelism, compute bottlenecks)

• Ability to reason about performance at multiple levels (kernel, model, system)

Nice to Have

• Experience optimizing machine learning or deep learning workloads

• Familiarity with video generation, diffusion models, or related domains

• Experience with compilers, runtime systems, or ML frameworks internals

• Background working with AI accelerators (e.g., GPUs, or custom silicon)

Send your CV to: Mika@decart.ai