texture image
texture image
0%

ABOUT

Hi! I'm Toluwani, a curiosity-powered human on a never-ending quest to understand human and artificial minds, why we make the choices we do, what freedom truly means, and how we can align intelligence with intention. I ask big questions, follow weird thoughts, and believe the most fascinating answers are the ones that lead to even better questions. If you're into brains, philosophy, or figuring out how AI can be less of a black box and more of a trustworthy companion, we'll get along just fine.

GOOGLE SCHOLAR
toluwani image

Current Exploration

I aim to build Safe and Trustworthy Generative AI Systems. Specifically, I am currently exploring (1) the security issues of watermarking in Generative AI and (2) the alignment problems that comes from the use of watermarks.

I am also working on and open to collaborations in various domains, including, but not limited to:

  • Philosophical Approaches to Alignment
  • Model Evaluation
  • Societal Impacts of AI
  • Self/Agentic Alignment
  • Threat Modeling and Security Games
  • Scalable Oversight

Feel free to reach out if you have an idea or project where we can collaborate!

Past Research Experience

As a student at MBZUAI, I served as a Research Assistant at the SPriNT-AI Lab, collaborating with Dr. Karthik Nandakumar. My primary focus was the development of an innovative method known as PolyKervNets, designed to address the Accuracy-Efficiency challenge within Deep Neural Networks, particularly in the context of private inference (PI).

Additionally, I embarked on personal projects covering Accented Speech Recognition, Racial Bias Mitigation in SSLs, and Gender Bias Mitigation in Word Embeddings. These projects not only broadened my understanding of AI but also contributed to bias mitigation and performance enhancement in various applications.

During my Applied Science Internship at G42, I played a pivotal role in developing a toolkit aimed at streamlining the preprocessing of the NHANES dataset, ultimately facilitating model utilization, result reproduction, and TRIPOD statement extraction.

Subsequently, as an AI research assistant at MBZUAI Metaverse Lab under the guidance of Prof. Abdulmotaleb El Saddik, I embarked on projects to enhance 2D classifiers in the Proactive Violence and Weapon Detection domain, including the leadership of an innovative project known as ScholarFace.

Education

My academic journey includes a Bachelor of Science in Computer Science from Adeleke University, Nigeria, and a subsequent Master of Science in Computer Science from the University of Ibadan, Nigeria.

More recently, I achieved a Research-focused Master of Science in Machine Learning at the Mohamed Bin Zayed University of Artificial Intelligence in December 2022.

Presently, I am pursuing a PhD in Machine Learning at MBZUAI, working under the guidance of Dr. Nils Lukas.

GALLERY

gallery
gallery
gallery
gallery
gallery
gallery
gallery

SOCIAL

I am open to collaborations and can be reached through