Hello! ๐ I’m Sungnam Park, working as a Tech Lead & Senior Machine Learning Engineer at Karrot (Daangn).
Introduction
I am a Machine Learning Engineer with expertise in large-scale recommendation systems, ranking, and applied deep learning. I have a strong background in ML research and production deployment, with experience in leading ML projects, mentoring, and optimizing models for user engagement and business impact. I am passionate about building AI-driven products that enhance user experience and drive business growth.
Experience
Karrot (Daangn) | Senior MLE & Tech Lead | 2022.04 - Present
- Tech Lead, Diversity & Serendipity Part, Feed Quality Team (2024.09 - Present)
- Led Diversity initiatives, increasing new service clicks by +20~25% through optimal exposure methods.
- Machine Learning Engineer, Feed Quality Team (2023.02 - 2024.09)
- Designed scalable retrieval pipeline, improving Home Feed engagement (+3~5% chats).
- Deployed ranking model for Short-form feed, improving CTR by +15~20%.
- Machine Learning Engineer, Recommendation Team (2022.04 - 2023.02)
- Built LLM-based pipeline extracting item attributes across global regions (KR, US, UK, JP).
- Improved Detail-page ranking model, increasing clicks by +2~5%.
Kakao | MLE & Tech Lead | 2020.02 - 2022.03
- Tech Lead, Research Part, Recommendation Team (2022.01 - 2022.03)
- Led two-stage deep learning recommendation POC to improve user engagement.
- Mentored interns and junior engineers.
- Machine Learning Engineer, Recommendation Team (2020.02 - 2022.01)
- Developed a two-tower DeepFM model, boosting KakaoTalk third tab CTR by 0~5%.
- Built a sequential recommendation model for Melon(music app), increasing click count and CTR by +10~15%.
- Implemented diversification methods, improving category coverage by +25~30% with a minor CTR drop.
- Maintained core recommendation libraries, optimizing text feature extraction.
Key Achievements
Challenges & Hackathons
- RecSys Challenge 2021: Achieved 8th place with advanced click prediction and target encoding techniques.
- NAVER AI Hackathon 2018: 1st place for end-to-end Korean speech recognition model using CNN + Seq2Seq.
Public Speaking
- Daangn Tech Meetup: Presented on designing a scalable Content-based Retrieval Model Pipeline (7+ app spaces, 1,000+ daily runs) for home feed and beyond.
- Google Cloud AI Summit Seoul ‘24: Speaker discussing Generative AI use cases at Karrot.
- PR12 Paper Reading Study: Presented deep learning papers on recommendation systems & NLP (Available on YouTube).
Technical Skills
Programming Languages / Environments
- Primary Languages: Python, Go
- Platforms: Linux, Mac OS
Expertise
- Large-scale recommendation systems
- Ranking algorithms
- Deep learning model development and deployment
- A/B testing and experiment design
- Machine learning pipeline construction
Education
Yonsei University | 2013.03 - 2020.06
- B.S. Information & Industrial Engineering
Contact
- ๐ง Email: sungnam1108@gmail.com
- ๐ฑ Phone: +82-10-4870-8573
- ๐ GitHub: angrypark
- ๐ผ LinkedIn: linkedin.com/in/angrypark
- ๐ Resume: resume.pdf
This page will be continuously updated.