Youngmin Ko
AI Forward Deployed Engineer at KRAFTON
At KRAFTON, I build products and features that combine domain rules and AI for meetings, customer support, finance, and business analysis. I lead meeting automation development and maintenance and contribute data validation, enterprise integrations, and operational recovery to business-owned advertising and feedback-analysis products. Earlier, I independently built anomaly detection for unmanned ground vehicle communication links without labeled faults during an eight-week internship at Hanwha Aerospace.
Current work
- Building products and features at KRAFTON that combine domain rules and AI for meetings, customer support, finance, and business analysis.
- Researching the reliability of RL reward and verification signals, and why time-series anomaly-detection metrics disagree.
- Working on an open-source toolkit for time-series anomaly-detection evaluation, this portfolio, and an archive of technical writing.
- A paper on replay-strategy selection is accepted at a NeurIPS 2026 workshop.
Links
- CV: mandu05.com/cv · GitHub · LinkedIn · Email: rhdudals0505@naver.com
- Another public page in this wiki: Core statement