Research

Multimodal AI for Oral Cancer Detection

At UCL, I am developing multimodal AI methods that combine genetic profiles and medical imaging to improve early detection of oral cancer. This work involves designing and training deep learning models that can fuse heterogeneous clinical and imaging data to support more accurate and earlier diagnosis.

Neuromorphic Vision and Event-Based Perception

My PhD research focused on advancing autonomous perception in robotics using neuromorphic vision systems. I have developed models for event-based segmentation and panoptic perception, combining RGB frames and event data with architectures such as SegNet, transformers, graph neural networks, and spiking neural networks.

Machine Learning in Real-World Systems

In industry roles as a Data Scientist, I have built and deployed machine learning models for computer vision, NLP, and risk modelling, including claim classification, motion prediction, and anomaly detection, with a focus on robust pipelines and production-grade code.

Publications

Below is a selection of my recent work. For a complete and up-to-date list, please see my Google Scholar profile: Google Scholar – full list

Preprints

Journal Articles

Selected Conference Proceedings

Other

For full citation details, co-author list, and links to PDFs where available, please refer to: Google Scholar – Sanket Kachole