About
Research
I'm a PhD student at the Institute of Machine Learning, University of Edinburgh, working on applying machine learning methods to understand how the brain processes information.
My research addresses a fundamental challenge in neuroscience: when we record from neural circuits, we observe a noisy mixture of external inputs, internal dynamics, and measurement error. How do we disentangle these components to understand what the circuit is actually computing?
I develop identifiable generative models to recover the latent inputs and internal dynamics shaping neural activity.Currently, this involves developing low-rank input decompositions for linear dynamical systems and control-theoretic approaches for nonlinear recurrent networks. More broadly, I'm interested in inverse problems: given high-dimensional measurements from a complex system, what underlying structure or process generated them?
Entrepreneurship & Applications
Beyond research, I'm interested in building things that matter. I'm drawn to entrepreneurship and applying ML methods to solve real problems, whether in neurotechnology, computational tools, or other domains where these techniques can create practical impact. I'm motivated by opportunities to work across different contexts and contribute to projects with real-world value, from early-stage startups to grassroots initiatives.
Projects
Recent work in computational neuroscience and machine learning
Neural Unmixing: Decoding Recurrent Dynamics and Structured Inputs
Implementation and evaluation of structured Multivariate Autoregressive (MVAR) and Latent Dynamical System (LDS) models with low-rank inputs for analyzing neural time series data. Developed tailored parameter estimation algorithms and a nested cross-validation framework to robustly recover latent dynamics and structured variability across trials. This work forms the foundation for my ongoing PhD research.
Disentangling Input-Driven and Intrinsic Neural Variability
Integration of two neural dynamics modeling frameworks, Low Tensor Rank RNNs and iLQR-VAE to analyze learning-induced changes in neural connectivity by inferring inputs and connectivity changes in nonlinear dynamical systems from neural data. Work involved wrestling with fundamental identifiability challenges in separating intrinsic dynamics from external forcing in non-linear systems. This work forms the foundation for my ongoing PhD research.
Multi-Label Safety Classification with Active Learning
Multi-label safety classification system for imbalanced datasets, integrating active learning with human annotator feedback. Fine-tuned transformer models (T5, LLaMA, BERT) using LoRA and 8-bit quantization, comparing generative multi-label (seq2seq) and discriminative binary classification approaches. Exploration of LLM-assisted annotation pipelines (LLMaAA methodology) combining prompt engineering with active learning strategies for scalable data collection.