Ilia Semenkov
PhD, Research Scientist (AI/ML)
I'm an AI Research Scientist with a recent focus on neuroimaging and biosignals, and additional experience in computer vision and multimodal representation learning. Across projects, I build models that connect brain recordings with naturalistic stimuli (language and audiovisual content), and I take ownership of technical execution from data pipelines through training, evaluation, and releases. I mentor students and junior researchers, and I am comfortable driving technical decisions within collaborations. I am interested in roles with more scope and ownership, where I can set direction, drive collaborations, and deliver research end-to-end.
Research themes
Brain decoding
ML for brain recordings (e.g., EEG/MEG/fMRI/ECoG) that links neural activity to stimuli, behavior, and task-relevant brain states.
Multimodal learning
Methods that connect brain data with language and audiovisual representations within shared embedding spaces.
Interpretable models
Models with explicit structure that makes it possible to inspect what drives predictions and how information is represented.
Closed-loop and real-time ML
Low-latency tracking and forecasting designed for feedback-in-the-loop settings such as neurofeedback and BCI.
Representation learning
Learning useful embeddings that transfer across tasks, support downstream modeling, and stay robust.
Computer vision
Work across core CV tasks, including recognition, segmentation, depth-based reasoning, and generative modeling.