Hands‑on experience with AI frameworks, including PyTorch, vLLM, SGLang, Unsloth, TensorRT‑LLM, Megatron‑LM, and DeepSpeed. Proven experience in LLMs, Generative AI models, transformer architectures, and end‑to‑end AI pipelines. Familiarity with AMD MI‑series GPU architecture, GPU kernel programming, and the ROCm AI software stack is strongly preferred. Strong communication and presentation skills, with the ability to articulate architectural proposals and value propositions clearly.