With the introduction of Model-as-a-Service, integrated SSO, and access to local VNG Cloud services, this release marks a major step forward in delivering a seamless AI development experience tailored for modern enterprises in Southeast Asia.

This is the story of Poptech - a forward-leaning tech company in Asia - fine-tuning Whisper Large V3 (10B) to make voice AI work for Southeast Asia—using GreenNode as their infrastructure partner.

Use case from a major robotics and automation company to train and fine-tune large-scale Large Language Models (LLMs), scale their operations seamlessly, and unlock a new generation of AI-powered service robots

This project reminded us that building AI Cluster infrastructure at scale isn't just a technical challenge — it’s a people challenge

Break down the core architectural components of an AI cluster — including node, rack, and cluster-level design — and how an in-house team like ours managed to bring everything together and make it actually work

As one of the first teams to build a large-scale AI cluster in Asia, we’re here to share our journey — the real, gritty, behind-the-scenes lessons learned. This is part of an ongoing blog series where we walk you through the entire process, from uncertainty to execution.

Artificial intelligence (AI) is becoming pervasive in the enterprise. To support these AI applications, businesses look toward optimizing AI servers and performance networks. Many enterprises prioritize GPU acceleration while overlooking a critical component: Storage infrastructure. According to Gartner, most enterprises will not have to build new storage infrastructure for generative artificial intelligence (GenAI) because they will fine-tune existing large language models (LLMs) instead of training new models. However, large-scale GenAI deployments will still require distinct storage performance and data management capabilities for the data ingestion, training, inference, and archiving stages of the GenAI workflow, rather than relying on traditional cloud storage.

ach NIM package includes pre-tuned models, APIs, and all runtime dependencies—optimized to serve models like DeepSeek, LLaMA, Mistral, and SDXL at blazing-fast token throughput, ready to plug into your app via REST or gRPC. Now with NIM on the GreenNode AI Platform, deploying AI inference is finally what it should’ve been all along: instant, scalable, and painless.

Starting with our own customer’s pain points in Slurm management, we eliminate these challenges with intelligent workload distribution and optimized resource management. Our Managed Slurm Cluster enables multi-GPU, multi-node training, ensuring that every compute cycle is maximized for efficiency.