Jimmy Shong
I am a first-year Master of Science in Computer Science (MSCS). I currently study ML systems under the guidance of Prof. Fan Lai . I am particularly focused on improving the efficiency and fault-tolerance of systems for pre-training large foundational models. I also had the great fortune working with Prof. Song Han at the MIT HAN lab as a research intern studying on-device ML. My other interests include ML runtime optimization and AI for Science.
Experimental results of instruct-tuning Llama3-8B using numerous supervised finetuning (SFT) methods with Llamafactory.
A 4.6M modified ResNet that achieves a 96.1% accuracy on CIFAR-10, a 3% increase over the reported accuracy of the 11.4M ResNet-18 model.
A web tool that allows website owners to scan their websites for vulnerabilities and connect with certified cybersecurity professionals.
A ML package was designed to analyze trajectories of nanodiamonds in macrophages and predict their diffusion type.