Biography

I am a final-year PhD student at the Safe Artificial Intelligence Lab, Department of Computing, Imperial College London. My research focuses on computer vision, robust training, and robustness validation through generative modeling. During my PhD, I had the opportunity to intern at Microsoft ASG, Amazon AGI, and Huawei Noah’s Ark Lab, where I worked on diverse projects on LLMs including:

• Distillation-based Quantization Aware Finetuning of language models with weight sharing,
• Video Understanding, Multimodal generation using Mixture-of-Experts (llama-moe), Spatial control of transformer based models (both autoregressive and diffusion),
• Reinforcement Learning based Text-to-Speech alignment.

Alongside my PhD, I also worked part-time as a Machine Learning Research and Software Engineer at Safe Intelligence Ltd. for 2 years. Before joining Imperial, I worked as a software developer at Daedalean AG, the coolest place to be at for working on certifiable autonomous autopilot for manned aerial vehicles.