Research Kernel Engineer (AI)

Location: Singapore or Austin, US

Company

hirify.global is a world-leading technology company specializing in Bitcoin mining solutions and AI cloud infrastructure.

What you will do

• Write and optimize CUDA and Triton kernels for new quantization schemes, sparsity patterns, and attention variants.
• Conduct performance attribution across the serving path using roofline analysis to identify dominating kernels.
• Convert published research methods into production-viable, high-performance implementations.
• Develop the measurement basis and profiling frameworks used by the research team.
• Collaborate closely with the AI Cloud platform team to integrate kernels into the serving stack.

Requirements

• Degree in Computer Science, Electrical Engineering, or a related field.
• Proficiency in CUDA, Triton, Python, and C++.
• Deep understanding of GPU architecture and memory hierarchy, including occupancy and tensor cores.
• Experience optimizing inference-critical paths such as attention, GEMM, and KV-cache management.
• Ability to implement complex methods from research papers or whiteboard sketches.
• Must be based in or able to work from Singapore or Austin, US

Nice to have

• Experience with inference engines like vLLM, SGLang, or TensorRT-LLM.
• Knowledge of compilers or IR-level tools such as MLIR, TVM, or TorchInductor.
• Experience with distributed serving, including tensor and pipeline parallelism.
• Publications at top-tier systems venues or significant open-source contributions to GPU computing.

Culture & Benefits

• Inclusive environment with an exciting startup spirit and open workspaces.
• Opportunity to network with industrial pioneers and contribute to the digital asset industry.
• High degree of personal accountability, autonomy, and fast growth opportunities.
• Attractive welfare benefits including training and mentoring.