Large language model (LLM) research at NUS


We are always looking for PhD students, postdocs, and visiting students to join our lab or collborate with us for research on Large language models (LLMs) at the National University of Singapore (NUS). Our projects aim to contribute to the development of next-generation Large Language Models (LLMs) and the better usage of the current LLMs. We consider new methods to improve the performance of LLMs while advancing the understanding of LLMs.

If you are interested in joining our lab or collaboration, please contact me at kenji at comp.nus.edu.sg
About PI of the lab
  • Name: Kenji Kawaguchi
  • Education: MIT / Harvard University​​​​​​​
  • Position: Presidential Young Professor
  • Publications: Google Scholar
  • Profile: homepage
  • Research interests in our lab at NUS:  Large Language Models (LLMs) and Deep Learning
ABOUT OUR LAB
Computational resources
We have sufficient computational resources to conduct research on LLMs. We are constantly updating the computational resources. Our lab has access to private two DGX servers with 16 A100 GPUs, one cluster with 8 L40S GPUs, 12 GPU serviers with 96 A4000 GPUs, and NSCC with more than 50 A100 GPUs. We also use Google Cloud Credits on top of these.

We can also purchase your own GPU workstation if needed. Before (as well as after) joining our lab, you can negotiate to purchase your own GPU workstation. NUS and NUS CS have shared GPU clusters, but it is convenient to have your own GPU workstation depending on your research style. We will try our best to avoid the situation where the computational resource is the bottleneck of our research, by accommodating the needs for computational resources as much as possible.

Publications
Our lab will aim to publish impactful works in top conferences in deep learning (NeurIPS, ICLR, ICML, CVPR, AAAI, etc,) as well as top journals in computational physics and applied mathematics.

Collaborations
To advance deep learning and the understanding of intelligence, our lab collaborates with researches at MIT, Harvard University, Brown University, Stanford University, Google, and other top institutes around the world.

Co-advising option
I am open to co-advising PhD students at NUS.

Lab members can also expect...
Flexibility in choosing research themes and topics.
Assistance with obtaining internships, postdoc positions and faculty positions.