Yusuf Hakan Kalaycı

Yusuf Hakan Kalaycı

Artificial Intelligence Research Advisor

As a Research Advisor at Palmate, Yusuf works at the intersection of theoretical computer science and artificial intelligence. He holds a PhD in Computer Science from the University of Southern California (USC), where his research centered on the algorithmic foundations of efficient and reliable AI. His recent work focuses on inference-time optimization for large language models - adaptive generate–rank–verify search and optimal-stopping strategies that preserve model quality while cutting computational cost, in one case matching standard best-of-N sampling with markedly fewer generations. His broader research spans machine learning theory, stochastic optimization, algorithm design, and computational social choice, with peer-reviewed papers at leading venues such as STOC, AAAI, and ICALP. The problems he has tackled range from PAC and transductive learning to stochastic packing, proportional representation, multiwinner voting, and long-term fairness in rotating citizens' panels. A gold medalist for Türkiye at the International Olympiad in Informatics, Yusuf pairs a competitive-programming pedigree with rigorous, internationally oriented research. At Palmate, he turns that depth into practical strategies for building more efficient, reliable, and scalable AI products.