About

I’m a postdoctoral researcher at Royal Holloway, University of London (Leverhulme Trust) and a part-time Research Associate at King’s College London. My work sits at the intersection of large language models, behavioural fine-tuning, and agent-based simulation.

At RHUL, I’m running large-scale studies benchmarking LLM-simulated behavioural archetypes against human participants across game-theoretic tasks, and building LoRA/PEFT fine-tuning pipelines to give small language models psychologically grounded behavioural personas. At KCL, I contribute computational modelling to interdisciplinary research on international dispute resolution and state compliance behaviour.

Previously, I was a Research Fellow at University College London, where I built an agent-based model to analyse and predict state compliance with European Court of Human Rights judgments - translating outputs into policy-relevant insights for legal and political science audiences.

I hold a PhD in Computer Science from the University of Edinburgh (2023), where my thesis focused on agent-based modelling for tobacco control policy, and an Integrated BS-MS in Physics from IISER Mohali.

Research interests: LLM behaviour and alignment, parameter-efficient fine-tuning, multi-agent simulation, computational social science, AI for policy.

Stack: Python · PyTorch · Hugging Face Transformers · PEFT/LoRA · Mesa · AWS · HPC