Backend/Infrastructure Engineer at Mubit — AI memory infrastructure startup

As the first engineering hire, you will architect Mubit’s core infrastructure from scratch. You’ll build high-performance systems for sub-80ms memory retrieval and shared state management, directly solving the "forgetful agent" problem. This role involves deep distributed systems work, SDK development, and close collaboration with the founding team to scale production AI agent fleets.

• Design and implement the core storage and retrieval architecture to handle agent execution history at massive scale.
• Build and maintain the SDK that developers use to integrate Mubit with frameworks like LangGraph, CrewAI, and AutoGen.
• Profile and optimize system bottlenecks to ensure consistent performance under heavy load while scaling infrastructure 100x.

• Proven experience building and running high-scale, latency-sensitive production systems using performance-oriented languages like Rust, Go, or C++.
• Deep understanding of distributed systems, storage engine trade-offs, and serving data within a hot execution loop.
• A self-directed builder who thrives in ambiguity and has extensive experience with LLM orchestration frameworks and agentic memory patterns.

• Founding engineer impact: As hire number one, you co-own the architecture and technical direction of a critical infrastructure layer for the AI era.
• High-performance engineering: Tackle complex distributed systems challenges where sub-80ms latency is the baseline requirement for production agents.
• On-site collaboration: Build alongside a well-funded founding team in a high-energy Canary Wharf office, fostering rapid iteration and deep technical ownership.