Alec Sargood
PhD AI Researcher @ UCL. Specialising in Generative Models and Computer Vision.
I am an Applied Mathematician and AI Researcher currently pursuing my PhD in Computer Science at University College London (UCL) within the Hawkes Institute.
My current research focuses on the theory of generative models and computer vision, with applications in medical imaging, data generation tasks, and inverse problems. Much of this work centres on diffusion and flow models, both in understanding their underlying dynamics and in adapting them to settings where high-quality data is scarce.
Open to collaboration: I am always keen to explore research collaborations in flow and generative models (both theoretical and practical), image and data synthesis, and general medical image problems. More recently, I have become interested in RL and SDE-based guidance techniques for the above, particularly controllable generation — steering text-to-image models, guiding data synthesis towards desired properties, and maximising rewards at inference time. Feel free to reach out!
Prior to my PhD, I completed an MRes at the University of Cambridge, an MSc in Mathematical Modelling and Scientific Computing from the University of Oxford, and a BSc in Mathematics from the University of Exeter.
news
| Mar 24, 2026 | I’m excited to share that I’ll be starting a 6-month PhD Research Internship at InstaDeep in London! I will be joining the team to focus on the guidance and steering of generative models! |
|---|---|
| Feb 20, 2026 | Thrilled to share that our paper GenTract: Generative Global Tractography has been accepted at CVPR 2026. |
| Nov 07, 2025 | Excited to share that CoCoLIT: ControlNet-Conditioned Latent Image Translation for MRI to Amyloid PET Synthesis has been accepted at AAAI 2026. This is joint first-author work with Lemuel Puglisi. |
selected publications
2026
- GenTract: Generative Global TractographyIn Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026Highlight Award
- CoCoLIT: ControlNet-Conditioned Latent Image Translation for MRI to Amyloid PET SynthesisIn Proceedings of the AAAI Conference on Artificial Intelligence, 2026* denotes joint first authorship (equal contribution).