vessl-ai

VESSL AI provides AI infrastructure and GPU-as-a-Service to help organizations secure necessary computing resources across self-serve and reserved cloud environments. The company builds a platform that supports training, fine-tuning, and serving for models ranging from large language models to multimodal foundation models, utilizing hardware such as NVIDIA A100, H100, H200, B200, B300, and GB300 chips. Its platform includes container compute workspaces, virtual machine clusters with InfiniBand networking, provisioned throughput APIs compatible with OpenAI and Anthropic, and centralized storage with team-level access controls.

Its user base spans enterprises, startups, and academic institutions, including organizations like Upstage, SqueezeBits, Holiday Robotics, UC Berkeley, Stanford, Subquadratic, and Nuance Labs. The company's work focuses on removing infrastructure bottlenecks so developers and researchers can manage AI workloads from development to production using per-second, usage-based billing and preconfigured environments.

The open roles at VESSL AI indicate that the company is building specialized research and engineering teams to advance agentic AI, reinforcement learning-based post-training, and distributed cache or accelerator kernel optimizations. The current openings include senior-level research roles, specifically seeking candidates with advanced degrees or significant practical experience in machine learning and systems. These positions are located in Seoul with remote work flexibility, focusing on core engineering challenges such as LLM inference efficiency, multi-node post-training pipelines, and translating research outcomes directly into platform products.