Second-year Computer Science student at UCL. I work on speech and audio AI: deepfake speech detection, spotting depression in spoken language, and how large language models represent real-world knowledge.
Looking for Summer 2027 internships.
Find out more about myself at my website:
Publications
- Enhancing Efficiency and Performance in Deepfake Audio Detection through Neuron-level Dropin & Neuroplasticity Mechanisms, IJCNN 2026 · arXiv · code
- Large Language Models for Depression Recognition in Spoken Language Integrating Psychological Knowledge, Frontiers in Computer Science, 2025 · paper · code
- and more publications before& in the future......
Projects
- Dropin: experiments for neuron-level Dropin on ASVspoof 2019 and FoR, with ResNet, GRNN and Wav2Vec 2.0 (Python, PyTorch)
- WeChat Emoji System: a small Windows tool for sending emoji in WeChat faster (C++)
- Hackathon Projects: Many hackathon projects including multi-agent workflow, using AI reconstruct gamified day, agentic system as coworker.
Open source
- WhatFills: A multi-platform tool to analyze disk usage and suggest clean up tactics
- kvplan: A commandline tool to plan vLLM deployments with/without the GPU, and find the most suitable model before downloading any.
- opencv/opencv#30597: fix a Python crash when creating stitching ImageFeatures objects
- apache/tvm-tirx#127: fix a divide-by-zero in iter map detection for predicates whose terms cancel
- and many other bugfix or tools......