Full Resume
Curriculum Vitae
Kangbeen Ko
Education
Advisor: Prof. SeungJun Kim
GIST, Gwangju, Korea
Skills
Programming Languages
Frameworks
Tools
Languages
Experience
Google Developer Group on Campus
Led AI/ML track as Core R&D member (2021–2022)
- Organized and taught sessions on Git, ML, and Kaggle studies
- Planned and hosted seminars, GDSC x GDG 'AI in Wonderland'
Mentored student AI projects (2022–Present)
- Delivered sessions on AI config management with Hydra/Gin
- Currently running 'Decoding The Deep' AI seminar (GenAI, RecSys, XAI, LLM fine-tuning)
- Provide ongoing student projects mentorship
Initial MVP development team for SickGPT, a chatbot for medical record understanding
- Designed a role-playing pipeline where agents interact to generate dialogue data, which was then used to fine-tune LLaMA2-7B.
- Developed an evaluation module measuring question relevance, BLEU score, and answer grounding
Authored a research paper on ML-based severity prediction for Chronic Kidney Disease
- Built a voice-based CKD severity prediction model
- Proposed clinically grounded features and validated statistical significance
- Published a research paper in eTELEMED 2024 (Advisor: Prof. Sejoong Kim)
Data-driven Hiring platform for Early-stage Startups
- Co-founded startup from 2021
- Raised Seed-round funding from two Venture Capitalists
- Built Personalized and Real-time Position Recommendation Algorithms using GNNs
- Led 2022 Tech Incubator Program for Startup (TIPS)
Projects
Soridam: Keeping Family Voices and Emotions Alive, Wherever You Are
2022Developed a family-oriented social networking app enabling communication through photos and voice messages across different regions and time zones. Integrated voice-based emotion recognition to preview messages with emotional emojis, and included a feature for revisiting memories by listening to archived voices of late family members.
Be With You: A Soft Hug for Every Silent Struggle
2024Developed a self-guided mental wellness platform powered by an LLM-based chatbot that provides psychological wellness framework to help users structure negative thoughts and ease emotional distress. With each response, comforting capybara characters gather around the user, culminating in a personalized message of encouragement generated by the LLM.
Peach Seoga: Text Simplification Platform for Slow Learners
2024-2025Collaborated with the non-profit Peach Market to develop an LLM-powered service that generates easy-to-read draft texts for learners with cognitive or developmental challenges. By uploading original content and selecting a desired format, users receive drafts based on prompts co-designed with experts — reducing the adaptation process from over a day to under one minute.
Awards
GIST Creative Convergence Competition 2nd Prize
Gwangju Institute of Science and Technology
Developed an AI model for forecasting solar power generation and optimizing energy scheduling on campus
Oasis Hackathon Special Awards
Honam ICT Innovation Square
A job aggregator that lets users search across all platforms and receive email for saved interests.
GIST Demo Day 2nd Prize
Gwangju Institute of Science and Technology
Developed a service enabling children to converse with their own drawn characters, featuring customizable TTS models built from parental voice recordings.
GDSC GIST Hackathon 2nd Prize
Gwangju Institute of Science and Technology
Developed a self-guided mental wellness platform powered by an LLM-based chatbot that provides psychological wellness framework to help users structure negative thoughts and ease emotional distress.
K-ium Medical AI Competition Excellence Award
Pusan National University Hospital
Built a deep learning model for cerebral aneurysm localization using cerebral angiography imaging data.
UN AI4Good Hackathon Chairman's Award, Innopolis Foundation
Ministry of Science and ICT, Korea
Developed a multimodal LLM-based online shopping assistant for visually impaired users, capable of extracting on-screen information and providing voice-based responses to user queries.
Other Experiences
Startup Program Finalist & Participant
Innopolis Foundation, KAIST E*5, X-IST, Ministry of Education
Selected for various early-stage startup competitions and accelerators, including the Pre-Startup Package (AI Track, Best Company Award), Youth Startup 300 (Finalist), Innopolis Campus (Final Selection), X-IST, and KAIST E*5 (Round 2).
Technical Book Reviewer & Beta Reader
Hanbit Media, BJ Public
Reviewed and contributed feedback for domestic and international IT publications. Authored endorsements and beta reviews for titles such as 'Reversed Python by Examples' and 'Reliable Data Engineering'; served as a preview reader for overseas titles in 2023.
Community Member – Developer Self-Growth Groups
Eigenvector, Geultto
Participated in developer self-growth communities focused on writing and technical development, including Eigenvector (3rd, 4th cohort) and Geultto (9th, 10th cohort).
English Language Proficiency
ETS / Seoul National University
TOEIC score: 935 / TEPS score: 426 — certified English proficiency for professional and academic use.
AI Talent Program Applicant
SGM AI
Successfully passed the document screening stage of the SGM AI 2nd cohort, a selective AI training program for aspiring researchers and engineers.
Challenge Program Leader & Member
GIST Infinite Challenge
Selected four times for the GIST Infinite Challenge program. Served as team member (5th cohort) and team leader (4th, 6th, and 8th cohorts), leading student innovation projects over multiple years.