🛠 Tech Stack

💼 About This Role

You'll design and operate the infrastructure for Krea's research and inference, including distributed training, GPU clusters, and petabyte-scale data pipelines. Your work will directly power next-generation AI creative tools. You'll build custom distributed datastores and streaming pipelines from scratch, solving complex orchestration and scaling challenges at massive scale.

🎯 What You'll Do

• Design and build distributed data pipelines for petabyte-scale datasets
• Manage and scale 1000+ GPU Kubernetes clusters for training and inference
• Profile and optimize distributed training infrastructure (NCCL, InfiniBand)
• Develop custom orchestration and fault tolerance systems for large-scale ML

📋 Requirements

• Strong fundamentals in distributed systems design and debugging
• Proficiency in Python and experience with data tools (DuckDB, PyArrow, Pandas)
• Experience with Kubernetes for container orchestration at scale
• Deep understanding of OS, networking, and file-system internals

✨ Nice to Have

• Experience with PyTorch internals and custom dataloaders
• Knowledge of NCCL, InfiniBand, or RDMA
• Experience with streaming systems like Kafka or Pulsar

🎁 Benefits & Perks

• 🏖️ Flexible PTO
• 🏥 Health insurance
• 📈 Equity