Iโm a passionate AI researcher and developer interested in deep learning, visionโlanguage models, and NLP applications.
I enjoy researching and developing systems in these fieldsโfrom training neural networks to deploying real-world applicationsโespecially those that combine perception and language understanding.
Dec. 2024 โ Jun. 2025
- Trained custom wake word detection models and converted them to ONNX format.
- Successfully embedded the ONNX wake word system into an Android real-time inference app.
- Developed and quantized YOLO-based ensemble models for vision tasks, then embedded them into Android apps.
- Worked on Human Detection, Face Recognition, Emotion Recognition, Action Recognition, and Gaze Estimation.
- Built and deployed a RAG system using Triton Inference Server, Llama 3, Milvus, and Redis.
Mar. 2025 โ Present
- Researching hallucination mitigation in multimodal LLMs (MLLMs) via contrastive decoding and RAG.
Mar. 2024 โ Feb. 2025
- Served as a core member and project lead.
- LG Aimers / Data Intelligence Internship (Jul. 2024 โ Aug. 2024): Developed data pipelines and exploratory analyses for real-world AI use cases.
Mar. 2023 โ Dec. 2024
- Backend developer for Silver Campus, an internet-lecture platform for seniors.
- Repository: SilverCampus/aws_server
Jan. 2023 โ Mar. 2023
- Established low-carbon building management guidelines based on building carbon-emission analysis.
Mar. 2020 โ Dec. 2020
- Participated in Kakao Impact X MYSC 100UP Problem-Definition Workshop (Jul. 2020 โ Aug. 2020).
- Helped secure a โฉ40 million grant for a collaborative clothing-manufacturer support program.
Feel free to reach out via GitHub or LinkedIn!
- ๐ LinkedIn: kyubin-kim
- ๐ง [email protected]
Thank you for visiting my GitHub profile! ๐