🛠 Tech Stack

💼 About This Role

You'll architect end-to-end search and retrieval platform for Twelve Labs' multimodal AI and set technical direction for production-scale systems worldwide. You'll co-design retrieval strategies with research and training teams, optimizing latency, recall, cost, and scalability. This role offers high autonomy and impact, working across search, ML serving, and distributed systems.

🎯 What You'll Do

• Architect search platform end-to-end on EKS
• Optimize million to billion-scale retrieval across vector and lexical paths
• Set engineering patterns for production microservices
• Define and own evaluation frameworks for search quality

📋 Requirements

• 8+ years building production ML systems with search or retrieval experience
• End-to-end ownership of complex systems from problem identification to delivery
• Strong Python skills; Go experience is a plus
• Hands-on with ML model serving (KServe, Triton, Ray Serve)

✨ Nice to Have

• Experience with multimodal or video search/retrieval systems
• Experience with ANN index tuning at billion-vector scale
• Background in end-to-end retrieval co-optimization

🎁 Benefits & Perks

• 🥑 Free snacks, coffee, and fresh food in office
• 💻 MacBook and 700K KRW home office setup renewed every 3 years
• 🍱 Monthly 600K KRW card for meals, transport, etc.
• 🏖️ 2-week winter break at year-end
• 🩺 Annual health checkup

📨 Hiring Process

Estimated timeline: 2-4 weeks

1. 1Application Review· N/A
2. 2Recruiter Interview (30 min, video)· 30 min
3. 3Loop Interview: Hiring Manager & Live Coding (90 min, onsite)· 90 min
4. 4Loop Interview: System Design & Final Round (90 min, video)· 90 min
5. 5Reference Check· N/A
6. 6Offer· N/A