Andrii Skliar

Andrii Skliar is a deep learning researcher at NVIDIA, where he develops methods and systems for efficient large language model inference. His work spans mixture-of-experts models, speculative decoding, and GPU kernel optimization, including work with vLLM and FlashInfer. His research connects model design with inference performance to make large language models faster and more efficient at scale. He holds an MSc in artificial intelligence from the University of Amsterdam.