Professional Experience

Summary

Undergraduate intern conducting in-depth research on explainable AI (XAI) and diffusion models with a focus on adversarial perturbations under Prof. Jaesik Choi.

Key Responsibilities

• Reviewed recent literature on XAI and adversarial machine learning
• Implemented and evaluated diffusion model variants with adversarial perturbation techniques
• Collaborated with research team for experiment design and result analysis
• +1 more responsibilities

Key Achievements

• Developed novel adversarial perturbation purification methods improving model robustness
• Contributed to ongoing research manuscripts for top AI conferences

Technologies Used

Skills Developed

Major Projects

• Adversarially Robust Diffusion Models: Designed and tested diffusion models with integrated adversarial perturbation to improve robustness and interpretability.

Summary

Participated in an intensive bootcamp focused on AI development, DevOps, and MLOps practices in cloud environments, including hands-on LLM application deployment.

Key Responsibilities

• Developed end-to-end AI applications for telecommunications scenarios
• Collaborated on LLM-based real-world AI projects
• Managed cloud deployments and CI/CD pipelines
• +1 more responsibilities

Key Achievements

• Successfully deployed production-ready AI models in cloud environments
• Led a project improving network optimization via reinforcement learning
• Recognized for technical excellence and teamwork

Technologies Used

Skills Developed

Major Projects

• ArtChemy: AI-based Art Appreciation Assistant for Visually Impaired: Integrated VLM-based image captioning models to generate detailed descriptions of artworks. Applied prompt engineering techniques to tailor explanations according to the flow of art appreciation. Continuously improved UI and conversational interfaces based on feedback from visually impaired users. Optimized image classification, metadata processing, and AWS S3 integration for efficient data management.

Summary

Led research on AI-driven network security focusing on website fingerprinting and encrypted traffic classification.

Key Responsibilities

• Reviewed literature on encrypted traffic analysis and fingerprinting
• Designed and implemented deep learning models for traffic classification
• Collected and processed large-scale network traffic datasets
• +2 more responsibilities

Key Achievements

• Published papers in IEEE Access and KIISC journals
• Developed a traffic classification model achieving state-of-the-art accuracy
• Created a large-scale encrypted traffic dataset with over 50,000 samples
• +1 more achievements

Technologies Used

Skills Developed

Major Projects

• Website Fingerprinting Defense System: Developed adversarial machine learning-based defense mechanisms against website fingerprinting attacks.