Biography

I’m a research scientist and CUDA C++ developer. Since 2014, I’ve dedicated my career to leveraging GPUs to accelerate data processing and data management. My work focuses on developing massively parallel GPU algorithms that address hard-to-parallelize problems, including text processing (e.g., tokenization and parsing of formats such as JSON and CSV) and compression (e.g., Deflate decompression, LZ77, and Huffman decoding), as well as high-performance implementations of core parallel primitives such as sorting, top-k selection, and finite-state transducer (FST) operations. My research has led to multiple patents and peer-reviewed publications in top-tier venues such as VLDB and SIGMOD.

I currently work at NVIDIA on the CCCL team, where I develop and maintain core GPU libraries, i.e., CUB, Thrust, and libcu++, that power workloads across AI, high-performance computing, and data analytics.

Before joining NVIDIA, I worked on autonomous systems at Argo AI and on GPU-accelerated data analytics at HEAVY.AI (formerly OmniSci), a Bay Area startup developing a GPU-powered analytics platform.

I earned my PhD in Computer Science from the Technical University of Munich, graduating summa cum laude.