texture image
texture image
0%
toluwani image

Seeking answers to the big questions!

It seems like I've gotten your attention! I'm Toluwani and I ask a lot of questions. I am also a final-year Machine Learning PhD Candidate at MBZUAI, where I work on Generative AI Security, and on the lookout for full-time positions in industry (Applied Science, AI Safety/Security Research, Quant - Private/Public Markets).

Currently advised by Dr. Nils Lukas at MBZUAI, my doctoral research centers on robust watermarking techniques for Generative AI, provenance, and attribution. I recently completed a research visit at the Centre for Frontier AI Research (CFAR), A*STAR, hosted by Dr. Jie Zhang. Previously, I worked with Dr. Karthik Nandakumar at the SPriNT-AI Lab developing PolyKervNets for efficient private inference, interned as an Applied Scientist at G42 building biomedical data preprocessing pipelines, and earned a patent for ScholarFace while working with Prof. Abdulmotaleb El Saddik at the MBZUAI Metaverse Lab. I hold an M.Sc. in Machine Learning from MBZUAI, an M.Sc. in Computer Science from the University of Ibadan, and a B.Sc. in Computer Science from Adeleke University.

🔬 Research & Focus Areas

  • AI Safety & Security: Generative AI watermarking, threat modeling, provenance, and adversarial robustness
  • Societal Impacts & Governance: AI risk analysis, policy, bias mitigation, and ethical alignment
  • Agentic Systems: Multi-agent alignment, scalable oversight, and autonomous decision-making
  • Applied Machine Learning: Private inference, efficient architectures, and quantitative applications

If you are interested in discussing research, exploring innovative ideas, or collaborating, feel free to reach out anytime!

Updates

July 2026

Our paper "Mitigating Watermark Forgery in Generative Models via Randomized Key Selection" will be presented at the ICML-Trustworthy AI for Good workshop in Seoul! Currently under review at NeurIPS 2026.

June 2026

Our paper, "MirrorCheck: Efficient Adversarial Defense for Vision-Language Models", got the distinguished paper award at CVPR-AdvML! 🏆

WAXAL-NET is live now! Finetuned Edge ASR Across 19 African Languages, evaluated against 51 open source models.

May 2026

I got a gold reviewer award from ICML.

Our paper, "Watermarking Should Be Treated as a Monitoring Primitive", is submitted to NeurIPS 2026!

April 2026

I am at the Center for Frontier AI Research, A*STAR Singapore for a 3-month research visit. Looking forward to collaborating with the amazing researchers here.

March 2026

Our paper, "MirrorCheck: Efficient Adversarial Defense for Vision-Language Models", has been accepted at CVPR (AdvML) 2026! Congrats to the team.

Our paper, "Robust Safety Monitoring of Language Models via Activation Watermarking", is submitted to COLM 2026!

February 2026

Our startup idea is incubated under Kora. Excited for this new chapter.

Publications

image for Watermarking Should Be Treated as a Monitoring Primitive

Watermarking Should Be Treated as a Monitoring Primitive

Toluwani Aremu, Nils Lukas, Jie Zhang

NeurIPS'262026
image for Robust Safety Monitoring of Language Models via Activation Watermarking

Robust Safety Monitoring of Language Models via Activation Watermarking

Toluwani Aremu, Daniil Ognev, Samuele Poppi, Nils Lukas

COLM'262026
image for Mitigating Watermark Forgery in Generative Models via Randomized Key Selection

Mitigating Watermark Forgery in Generative Models via Randomized Key Selection

Toluwani Aremu, Noor Hussein, Munachiso Nwadike, Samuele Poppi, Jie Zhang, Karthik Nandakumar, Neil Gong, Nils Lukas

NeurIPS'262026
image for Aligning Stuttered-Speech Research with End-User Needs: Review, Survey, and Guidelines

Aligning Stuttered-Speech Research with End-User Needs: Review, Survey, and Guidelines

Hawau Toyin, Mutiah Apampa, Toluwani Aremu, Humaid Alblooshi, A. Valente, G. Leal, Z. Yue, Z. Talat, Hanan Aldarmaki

arXiv2026
image for MirrorCheck: Efficient Adversarial Defense for Vision-Language Models

MirrorCheck: Efficient Adversarial Defense for Vision-Language Models

Samar Fares*, Klea Ziu*, Toluwani Aremu*, Nikita Durasov, Martin Takáč, Pascal Fua, Karthik Nandakumar, Ivan Laptev

CVPR🏆2026
image for Optimizing Adaptive Attacks against Watermarks for Language Models

Optimizing Adaptive Attacks against Watermarks for Language Models

Abdulrahman Diaa, Toluwani Aremu, Nils Lukas

ICML🏆2025
image for On the reliability of Large Language Models to misinformed and demographically informed prompts

On the reliability of Large Language Models to misinformed and demographically informed prompts

Toluwani Aremu, Kemi Akinwehinmi, Emeka Nwagu, Syed Ahmed, Rita Orji, P. Arnau Del Amo, Abdulmotaleb El Saddik

AAAI2025
image for Library-Like Behavior In Language Models is Enhanced by Self-Referencing Causal Cycles

Library-Like Behavior In Language Models is Enhanced by Self-Referencing Causal Cycles

Munachiso Nwadike, Zangir Iklassov, Toluwani Aremu, T. Hiraoka, B. Heinzerling, V. Bojkovic, H. Alqaubeh, Martin Takáč, Kentaro Inui

ACL (Main)2025
image for All Languages Matter: Evaluating LMMs on Culturally Diverse 100 Languages

All Languages Matter: Evaluating LMMs on Culturally Diverse 100 Languages

Ashmal Vayani, Dinura Dissanayake, Hasindri Watawana, Noor Hassan, ..., Toluwani Aremu, et al.

CVPR🏆2025
image for A Call for Patience and Perspective

A Call for Patience and Perspective

Toluwani Aremu

X (Social)2025
image for Human Social Mirrors and AI Looking Glass

Human Social Mirrors and AI Looking Glass

Toluwani Aremu

Towards AI2025
image for The Thinker vs. The Builder

The Thinker vs. The Builder

Toluwani Aremu

X (Social)2025
image for Slow down!

Slow down!

Toluwani Aremu

X (Social)2025
image for ScholarFace: Scanning Faces, Discovering Minds

ScholarFace: Scanning Faces, Discovering Minds

Wong Yu Kang, Toluwani Aremu, Younes Balah, Maryam Nadeem, Ivo Gollini Navarrete, Abdulmotaleb El Saddik

Patent2025
image for Redefining Contributions: Shapley-Driven Federated Learning

Redefining Contributions: Shapley-Driven Federated Learning

Nurbek Tastan, Samar Fares, Toluwani Aremu, Samuel Horvath, Karthik Nandakumar

IJCAI2024
image for SSIVD-Net: A Novel Salient Super Image Classification & Detection Technique for Weaponized Violence

SSIVD-Net: A Novel Salient Super Image Classification & Detection Technique for Weaponized Violence

Toluwani Aremu, Li Zhiyuan, Reem Alameeri, Mustaqeem Khan, Abdulmotaleb El Saddik

SAI2024
image for Learning Rate Schedulers

Learning Rate Schedulers

Toluwani Aremu

Medium2023
image for Choosing a Learning Rate for DNNs

Choosing a Learning Rate for DNNs

Toluwani Aremu

Medium2022
image for What AI Still Can’t Do #003

What AI Still Can’t Do #003

Toluwani Aremu

Medium2022
image for What AI still can’t do #002

What AI still can’t do #002

Toluwani Aremu

Medium2022
image for What AI still can’t do #001

What AI still can’t do #001

Toluwani Aremu

Medium2022

Let’s work together

Got a cool research idea or just want to muse about the future? Hit me up!