Turkish Engineer Koray Kavukcuoglu to Lead Google DeepMind Operations

    Turkish engineer Koray Kavukcuoglu appointed to lead Google DeepMind operations
    Koray Kavukcuoglu will oversee Gemini development, advanced AI research and product teams at Google DeepMind.

    The longtime DeepMind executive will oversee Gemini models, advanced AI research and product teams as Demis Hassabis moves into a strategic science role

    By Bosphorus News Economy Desk

    Turkish engineer Koray Kavukcuoglu will lead Google DeepMind's operations, overseeing Gemini model development, advanced artificial intelligence research and the teams behind Google's flagship AI products.

    Kavukcuoglu will serve as senior vice president of Google DeepMind and report directly to Google and Alphabet Chief Executive Officer Sundar Pichai. He will retain his position as Google's Chief AI Architect, extending his influence across the company's research, infrastructure and product operations.

    Google stopped short of appointing him chief executive. The new structure separates operational control from the long-term scientific role assigned to DeepMind co-founder Demis Hassabis, who is leaving daily management to become chair of Google DeepMind and Chief Scientist of Alphabet.

    Hassabis will continue advising the company's model and research teams while concentrating on artificial general intelligence, scientific strategy and the societal consequences of increasingly capable AI systems. He will also remain chief executive of Isomorphic Labs, the AI drug discovery company created from DeepMind.

    Kavukcuoglu takes control of Gemini development

    Pichai said Kavukcuoglu would oversee Gemini model development, frontier AI research, the Gemini application and Google's AI developer teams. The remit places the research and product sides of the company's main AI operation under one executive.

    Kavukcuoglu joined DeepMind during its early years and has spent 13 years at the company, where he established its deep learning team and worked on projects including Deep Q-Network and WaveNet. He later became DeepMind's chief technology officer and took a leading role in coordinating research with Google Cloud and other commercial units.

    The new role expands responsibilities Kavukcuoglu had already taken on as Google sought to move model research into consumer, developer and enterprise products more quickly.

    Google said the Gemini application had passed 950 million monthly users, while demand for its models continued to rise among businesses and software developers. The leadership change comes during a demanding period for DeepMind, with Google under pressure to release new models faster and defend its position against OpenAI, Anthropic and lower-cost Chinese competitors.

    Alphabet shares fell about 4 percent after the restructuring was announced.

    Senior researchers leave to form Discovery Loop

    The management overhaul coincides with the departure of several of Google's best-known artificial intelligence researchers.

    Jeff Dean, Sanjay Ghemawat, Oriol Vinyals and Quoc Le are leaving to establish Discovery Loop, an independent public benefit corporation focused on machine learning, science and engineering. Google will invest in the company, provide cloud services and continue research cooperation.

    Dean and Ghemawat played central roles in building Google's early computing infrastructure, while Vinyals and Le contributed to research that helped establish the technical foundations of modern generative AI.

    Their departure removes several senior figures as Google tries to accelerate Gemini development and recover from delays affecting the next flagship version of the model.

    Kavukcuoglu inherits a research organization backed by Google's computing infrastructure and product network, but must manage the loss of senior researchers while keeping Gemini releases on schedule.

    From ODTÜ aerospace engineering to DeepMind

    Kavukcuoglu completed his undergraduate and graduate education in aerospace engineering at Middle East Technical University in Ankara before moving into computer science and artificial intelligence.

    He continued his studies at New York University, where he completed master's and doctoral work under computer scientist Yann LeCun. His 2011 doctoral thesis examined deep learning systems for object recognition, including methods for training hierarchical feature extractors.

    Before joining DeepMind, he worked as a researcher at NEC Laboratories America. His later work covered deep learning, generative models, unsupervised learning and reinforcement learning, with contributions to systems used in image recognition, audio generation and AI agents.

    The Royal Academy of Engineering elected him as a fellow in 2022, citing his technical work and leadership in applying artificial intelligence to scientific and engineering problems.

    Hassabis moves toward AGI and science

    Hassabis described the change as a decision to step away from operational responsibilities and concentrate on the development and governance of artificial general intelligence.

    He said increasingly capable AI systems required more attention to long-term scientific direction and their effects on society. His new role leaves him connected to DeepMind's senior leadership without responsibility for daily management.

    The division of responsibilities gives Kavukcuoglu authority over execution while preserving Hassabis's influence over research strategy. It also moves Google DeepMind closer to a structure in which model development, applications and commercial deployment are managed through the same chain of command.

    The restructuring places Gemini research, model development, applications and developer products under Kavukcuoglu as Google accelerates its competition with OpenAI and Anthropic.

    Sources: Google and Alphabet, Google DeepMind, Reuters, Middle East Technical University, New York University, Royal Academy of Engineering, Bosphorus News review and reporting.

    Home