Member of Technical Staff - Distributed Systems

The role

You will build systems that schedule, route, and coordinate AI workloads across Gimlet's infrastructure. You will work across scheduling, orchestration, control planes, APIs, and fault tolerance to enable workloads to run reliably across a heterogeneous fleet.

What you will do

• Build scheduling and orchestration systems for heterogeneous compute
• Design systems that manage independently scalable stages of distributed inference pipelines
• Improve the reliability and fault tolerance of production AI infrastructure
• Develop control planes and APIs that simplify how workloads are deployed and managed
• Improve resource management and scheduling as Gimlet expands across new accelerator types, nodes, and data centers
• Help the platform scale across additional hardware, nodes, and data centers

Required qualifications

• Experience building or operating distributed systems in production
• Strong software-engineering and systems fundamentals
• The ability to reason about concurrency, consistency, failure modes, and system tradeoffs
• Experience with scheduling, resource management, RPC, or asynchronous messaging
• A bachelor’s degree in a relevant field or equivalent practical experience

Nice to have

• Experience with Kubernetes or Kubernetes-adjacent systems beyond basic usage
• Experience designing service-oriented architectures using RPC or asynchronous messaging
• Familiarity with scheduling, queues, or resource management systems
• Experience building reliable APIs and operating systems under high load
• Software development experience in languages commonly used for systems development (e.g., Go, C++, Python)

Before you apply

• Gimlet Labs does not accept unsolicited resumes from recruitment agencies or search firms.

About the company

Gimlet is building the first multi-silicon neocloud designed for fast, efficient AI inference. They combine large-scale compute infrastructure with an execution platform that partitions AI workloads and maps each stage to the hardware best suited to run it.