I am an AI researcher and engineer working on multimodal and generative machine learning, with a particular interest in systems that have to work under real constraints: limited compute, scarce data, low-resource languages and real-time interaction.
I trained first as an engineer, completing a BSc (Hons) in Electronic and Electrical Engineering at Obafemi Awolowo University in Nigeria, where I graduated in the top five percent of my class and later served as a graduate assistant teaching microelectronics. In 2023 I joined the AI for Science master’s programme at the African Institute for Mathematical Sciences (AIMS) South Africa, affiliated with the University of Cape Town, as a Google DeepMind Scholar. I graduated in 2024 with Overall Distinction and was named Salutatorian with an Academic Excellence Award. My thesis, developed in collaboration with the University of the Western Cape, generated radiology reports from 3D chest CT scans by adapting a RadImageNet-pretrained ResNet-50 to volumetric data and decoding with GPT-2 through cross-attention, all on a single 24 GB GPU and a fraction of the CT-RATE dataset. The work was published by Springer Nature in the PAAISS 2025 proceedings and received the conference’s 2nd Best Paper Award.
As a Young African AI Research Fellow at InstaDeep (2024–2025) I moved into scientific machine learning for Earth observation, co-developing a physics-informed Transformer-VAE that inverts the PROSAIL radiative transfer model to retrieve vegetation properties from Sentinel-2 imagery. Keeping the physics inside the network gave interpretable latent variables and calibrated uncertainty, and convinced me that structure matters as much as scale.
Since June 2025 I have been an AI Research Engineer at IQAI, where I build real-time speech-to-speech pipelines and control systems for an embodied humanoid AI. I also co-authored a vision paper on large language models as operators of 6G radio access networks, presented at the NeurIPS 2025 AI4NextG workshop, and co-presented a tutorial on the mathematics of neural networks at the Deep Learning Indaba 2026 in Lagos. I am a member of Black in AI and am building NLP resources for Yoruba, my native language.
I am now seeking to pursue these questions through doctoral research: building multimodal models that combine evidence across modalities, respect known structure, and remain trustworthy in medical, scientific and everyday settings.