Notice
•
Application Deadline: May 10, 2026, 23:59 KST.
•
Research Participation Period: June 29, 2026 - August 21, 2026.
•
Location: Hybrid (on-site at KAIST with some periods conducted remotely)
We are looking for undergraduate students to join summer research participation program!
Note: For international students, only those currently enrolled at KAIST are eligible to participate in research activities, as we are unable to provide visa support for non-KAIST applicants.
Research at AMILab
Our ultimate goal is to develop socially intelligent machines. To achieve this, we are currently focusing on addressing challenges in machine perception and understanding, including multi-modal learning, video understanding, 3D perception, agentic system, reasoning in large language and multimodal language models (LLMs and MLLMs), and embodied intelligence with physical world interaction. We are also working on data-efficient learning to tackle the difficulties of acquiring large-scale real-world data.
If you are interested in AMILab, please refer to:
Research Activities
•
Research: You will work in a team with AMILab graduate students to explore and develop technologies needed in the current research community around a specific research topic.
We strongly encourage undergraduate students to pursue research publications during this program.
•
Regular Meetings and Seminars: You will participate in regular meetings and discussions to share research progress, as well as attend weekly seminars.
Research Topics for Summer Research Program
(topics may be adjusted based on research alignment and interest)
•
Audio and Audio-Visual Learning
◦
Full-duplex speech modeling and evaluation
◦
Causal understanding in ego-centric audio-visual settings
•
Agentic Systems
◦
LLM/VLM interface optimization
◦
Agent-native representations for graphical user interfaces
•
Reasoning in LLMs and MLLMs
◦
Spatial reasoning with MLLMs
◦
Efficient training for visual reasoning
•
Embodied Intelligence and Physical World Interaction
◦
Physically grounded 3D generation and correction
◦
Task progress assessment and failure recovery in embodied agents
Preferred Qualifications
•
Experience developing machine learning models, with proficiency in Python and PyTorch.
•
Strong interest in artificial intelligence and computer vision.
•
Self-motivated and proactive in pursuing research goals.
•
This program will follow a hybrid format, with certain periods conducted remotely and others requiring on-site participation.
•
Stipend support is not guaranteed but may be discussed upon request. If you’re interested, please contact us.
How to Apply
•
Please click “Application Form” button and fill out the information. After submitting the form, please send an email to kaist-ami-lab-admission@googlegroups.com.
•
If you have any questions, feel free to contact us: kaist-ami-lab-admission@googlegroups.com