Sundong Kim

I am an assistant professor at GIST AI Graduate School. My group focuses on developing human-like AI with abstraction and reasoning capabilities. We employ the following methodologies:

• Program synthesis
• Analogical reasoning
• Reinforcement learning
• Neuro-symbolic learning
• Representation learning
• Meta learning
• LLM agents
• Human-AI alignment
• Evaluation frameworks

As a mentor, I guide my students in proposing methodologies to address complex problems. I am interested in the methods themselves, helping refine experimental settings and discussing results through the lens of data science expertise.

E-mail: sundong (at) gist.ac.kr
Call: (+82)-62-715-6387
Office: AI Graduate School (S7) Room 204, GIST

• We are seeking collaborators with strong publication records (such as ICML, NeurIPS, ICLR, AAAI) to work with us in tackling visual abductive reasoning and creative math problems. We have a particular focus on reinforcement learning, neuro-symbolic program synthesis, and Human-AI alignment. This is an opportunity to work alongside passionate students in our group. Please contact me through this form.

---