Establish LLM,VLM,and audio model learning strategies and development roadmaps.
Build and operate data collection,curation,synthesis,and validation pipelines.
Conduct model training using SFT,RL,and continual training methodologies.
What they want
4+ years of hands-on AI/ML engineering experience.
Direct experience training and improving LLM or multimodal models.
Practical understanding and hands-on work with SFT,RL,continual training methods.
Nice to have
Pre-training or post-training experience with large language or multimodal models.
Built agent learning data for trajectories,tool calling,preference data.
Worked with image,video,audio data pipelines and model training.
Seoulstart's read
Visa tier (likely)
E-7 specialty / professional (likely)
Korean language
Business-level Korean required
English friendliness
Medium: English JD, mixed-language workplace
Context
Early-stage AI safety startup with strong VC backing (Samsung,Mirae Asset,Smilegate); Korean fluency necessary for daily collaboration and cross-functional alignment,but technical work conducted in English; stock options and international conference support signal growth-stage priorities.
This is Seoulstart's analysis of the public job description, not the employer's stated policy. Verify visa sponsorship, language requirements, and remote allowances directly with the recruiter.