Sailplane Tech Stack
Hierarchical planning AI for autonomous agent execution in production
Sailplane is a 2–10-person AI research company founded by Google Brain and Cloud veterans, focused on hierarchical planning for autonomous agents in real production environments. The tech stack (Claude, OpenAI, Python, TypeScript, Next.js, GCP) is optimized for rapid iteration on language models and agent systems. The hiring profile—five senior/principal engineers and one designer with accelerating velocity—reflects a research org scaling toward productization: projects span control planes, sandboxing, LLM deployment pipelines, and observability, while internal pain points center on managing large-scale AI data centers, indicating they're building both for external customers and their own infrastructure needs.
Tech Stack 12 technologies
What Sailplane Is Building
◆Challenges
- Autonomous management of massive ai data centers
- Managing massive ai data centers
▲Active Projects
- Control plane for autonomous agents
- Sandboxing and rollback strategies
- Llm production deployment
- Mlops pipeline implementation
- Performance benchmarking
- Llm deployment and mlops for on-premises customers
- Mlops pipeline for on-prem llms
- Agent stack development
- Simulation and evaluation pipelines
- Rich traces and observability
