Postdoctoral Researcher Position
Machine Learning Laboratory (MLLab)
Seoul National University (SNU)
Reimagining Efficient and Scalable Foundation Models
Prof. Hyun Oh Song invites applications for exceptional Postdoctoral Researchers interested in pushing the frontiers of modern machine learning systems. MLLab focuses on building the next generation of efficient, scalable, and deployable foundation models, spanning:
- Training-inference co-design for large models
- KV cache compression, quantization, sparsity, and memory-efficient LLMs
- Alternatives to standard Transformer attention (SSMs, memory architectures)
- Long-context and sequence modeling
- Model-based RL and world models
- High-stakes ML deployment
We are particularly interested in researchers who want to move beyond incremental improvements and instead rethink how large-scale models are trained, compressed, and deployed.
Who We Are Looking For
We seek candidates with:
- A Ph.D. in ML/AI or related fields
- Strong first-author publications at top-tier venues (e.g., NeurIPS, ICML, ICLR, CVPR, ACL, OSDI)
- Demonstrated ability to lead ambitious research projects
- Strong implementation skills in large-scale training (PyTorch/JAX, distributed systems)
- LLM efficiency or architecture design
- Long-context modeling or SSM-style models
- Reinforcement learning / world models
- ML for systems or hardware-efficient training
What We Offer
- Significant GPU resources for large-scale experimentation
- Freedom to lead independent research directions
- Opportunities to co-advise Ph.D. students
- Strong support for building an academic job market profile
Duration: 1-2 years (renewable)
Start date: Flexible
Salary: Competitive and commensurate with experience
How to Apply
Email hyunoh@snu.ac.kr
Subject: Postdoctoral Researcher Application - [Your Name]
Please include:
- 1-2 page research statement (vision + fit + start date)
- CV with publication list
- Two representative papers
- Contact information for at least two references (e.g. previous research supervisors and collaborators)
Applications will be reviewed on a rolling basis until filled.