My studies are in biomedical engineering, where I pursued Synthetic Biology, Genetic Engineering & Lab Automation. My passion was first in biology, then in computers and now I work on both as a Machine Learning Engineer at GSK. Fundamentally, my passion has always been understanding systems and systems dynamics, this is a perspective I applied for many years to understanding biology, but that is applicable to computers, agents and society. Currently, I am fortunate to be working on sequence learning transformer-based models to predict the structure and function of small molecules and nucleic acids for hyper-individualised drug development.

Brain-machine interfaces, human+AI design, systems biology, drug discovery, agents, UX in AI age (beyond chatbots, text interfaces or even screens)

at work: - end-to-end building of models an RNA sequence including the design of large experimental datasets to ML models that can reproduce experiments in-silico - highly ergonomic cloud tooling in my spare time: - open source package to organise UK's geographical data (boring but immensely valuable to anyone working with geo data) - full stack app for free access to etymological data (linking an english word to its roots in Latin and Indoeuropean) - ... more!

Born in China and based in the UK for the past decade, I’ve been passionate about AI since deploying my first computer vision model (a fine-tuned ResNet). I'm a curious and driven learner, with experience building AI systems across domains like marketing and consumer apps. Lately, I’ve been exploring the frontiers of generative AI—tinkering with multi-agent systems, personalisation, and open-data tools to push what's possible.

I'm keen to explore how AI can drive social impact, particularly in justice systems and the nonprofit sector—areas where data is rich but often underused. I'm also deeply interested in generative AI for personalisation, especially in creating adaptive, user-aware experiences in digital products. I’m looking to connect with collaborators working on responsible AI, civic tech, or mission-driven applications of GenAI, and would love to exchange ideas or build together.

At Trainline, I’ve worked on a Travel Assistant—an LLM-powered chatbot using a multi-agent framework to help travellers post-booking. I’m now exploring reinforcement learning with LLMs to personalise experiences across app and web. Outside of work, I’m building models to simulate the criminal justice system, exploring system dynamics and policy impacts. I’m also developing an open data AI platform to help nonprofits better use their data for social good.