I am an AI Architect on IBM’s Core AI team, where I am conducting research and building an AgentOps product that makes agentic AI systems reliable, observable, and governable at enterprise scale. My work spans framework design, evaluation tooling, optimization pipelines, and applied research for agentic systems.
I completed my PhD at the University of Edinburgh under Dr. Matthias Hennig and later conducted a postdoctoral fellowship at the Zuckerman Institute at Columbia University with Dr. Liam Paninski. I was also a member of the International Brain Laboratory (IBL) and the NSF AI Institute for Artificial and Natural Intelligence (ARNI). My research focused on building scalable machine learning models that connected neural activity to behavior at single-cell and single-spike resolution, spanning multimodal transformers, contrastive/self-supervised learning, and probabilistic latent-variable/state-space modeling. I also co-created SpikeInterface, a widely used open-source framework for flexible and robust spike-sorting pipelines.
Recent News
- I have joined IBM as an AI Architect to build an AgentOps platform for observing, evaluating, and optimizing enterprise AI agents.
- I will be presenting my work at the Bernstein Conference at a workshop on neurobehavioral modeling.
- I am co-organizing the workshop Foundation Models for the Brain and Body at NeurIPS 2025 in San Diego, United States.
- I will be presenting my work at the CTCN Computational Neuroscience Next Generation Symposium.
- I am giving a talk on “Jointly modeling neural activity and behavior using multimodal transformers” at the Bernstein Conference workshop, Toward a joint definition of neural-behavioral states.
- I co-organized the workshop Building a foundation model for the brain at Cosyne 2025 in Mont Tremblant, Canada.
- Our work Neural Encoding and Decoding at Scale was accepted at ICML 2025 as a spotlight.
- Our work In vivo cell-type and brain region classification via multimodal contrastive learning was accepted at ICLR 2025 as a spotlight.
- Our work Towards a “universal translator” for neural dynamics at single-cell, single-spike resolution was accepted at NeurIPS 2024.