Kian Mohadjerin

I am Head of AI at ElioVP and Founder of Wolker. At ElioVP, I work on AI infrastructure, AI Agents, GPU benchmarking, and high-performance inference deployment. Separately, through Wolker, I am building Paiton: an AI inference optimization stack focused on AMD GPUs. Paiton compiles and packages optimized AI models into deployable .so libraries, with compatibility for inference engines such as vLLM. Building AI Agents and vision models. My work focuses on making LLM and diffusion model inference faster, cheaper, and easier to deploy on AMD hardware. This includes ROCm/HIP kernel optimization, operator fusion, GEMM optimization, vLLM compatibility, and compiled model deployment through packaged .so libraries. I am interested in connecting with teams deploying AI inference workloads, especially those evaluating or scaling AMD or Nvidia GPU infrastructure.