LLM Model Compression & Acceleration PhD Position

About the role

Rakuten Asia, in partnership with Singapore's Economic Development Board (EDB) through the Industrial Postgraduate Programme (IPP), is hiring a fresh PhD candidate to work on large‑language‑model (LLM) compression and acceleration. The role combines academic research with practical engineering to make state‑of‑the‑art AI models more efficient for real‑time deployment.

Key responsibilities

• Design and implement advanced quantization and pruning techniques for LLMs, multimodal LLMs and diffusion models.
• Conduct large‑scale experiments on shared GPU clusters to evaluate memory and speed improvements.
• Collaborate with Rakuten researchers and EDB mentors on joint publications and open‑source tool development.
• Produce prototypes that demonstrate accelerated inference in real‑world applications.

Required profile

• Recent graduate (BSc/MSc) with a strong foundation in deep learning, machine learning and natural language processing.
• Demonstrated hands‑on experience building and training neural networks.
• Motivation to pursue a PhD while contributing to industry‑focused research.

Required skills

• Proficiency with PyTorch.
• Understanding of deep learning concepts.
• Knowledge of machine learning algorithms.
• Familiarity with natural language processing techniques.

What we offer

• Full sponsorship for PhD studies under the IPP scheme.
• Permanent employment with Rakuten Asia upon successful completion.
• Access to high‑performance computing resources and collaborative research environment.
• Opportunities for high‑impact publications, open‑source contributions and prototype development.