🛠 Tech Stack

💼 About This Role

You'll set the technical direction for TwelveLabs' next-generation multimodal embedding models, owning end-to-end model development from research to production. You'll drive large-scale distributed training and data strategy to deliver models serving thousands of companies worldwide. This high-autonomy role uniquely blends research, systems design, and cross-team leadership.

🎯 What You'll Do

• Set technical direction for multimodal embedding model architecture and training methodology
• Own end-to-end model development from research planning to production evaluation
• Architect and optimize distributed training infrastructure and GPU utilization
• Drive data strategy: design large-scale data curation and quality frameworks
• Define evaluation methodology and quality standards for embedding models

📋 Requirements

• 7+ years of industry experience in computer vision, NLP, or multimodal learning
• Deep expertise in large-scale distributed model training (FSDP, kernel optimization)
• Strong experience in contrastive learning or representation learning
• Strong proficiency in Python and PyTorch

✨ Nice to Have

• Experience training models at billion-parameter scale
• Experience with temporal video understanding or multimodal video modeling
• Track record of technical leadership in production ML environments

🎁 Benefits & Perks

• 💻 Latest MacBook and 700,000 KRW home office setup, refreshed every 3 years
• 🍱 Monthly 600,000 KRW corporate card for meals and transportation
• 🍿 Office snack bar with snacks, coffee, and fresh food
• 🌍 Global team working with B2B customers worldwide
• 🏢 Hybrid work model with autonomy and collaboration

📨 Hiring Process

Estimated timeline: 3-5 weeks

1. 1Application Review· 1 week
2. 2Recruiter Interview· 30 min
3. 3Loop Interview (Hiring Manager & Live Coding)· 90 min
4. 4Loop Interview (System Design & Final Round)· 90 min
5. 5Reference Check & Offer· 1 week