Software Engineer, Agent Infrastructure

Sapiom • San Francisco, CA

7d

About The Position

Sapiom is seeking a Software Engineer to join their Agent Infrastructure team. This role involves working on the core platform that enables agentic products to ship and scale. The company provides the necessary components for agents to act in the world, including compute, sandboxes, memory, identity, spend controls, storage, queues, and monitoring. The engineering team is described as world-class with deep infrastructure and payments DNA. The company has recently raised a $35M Series A led by Dragonfly, bringing their total funding to $50M. The engineering team operates with a flat organizational structure where engineers contribute across various problem domains as needed, reflecting the early stage of the company and the evolving architecture. Sapiom hires at two levels: engineers with one to three years of experience and Staff engineers. This structure allows for direct interaction with those setting the architecture, providing significant learning opportunities and the chance to have design decisions challenged by experienced engineers.

Requirements

• Strong engineering fundamentals and ability to write good code quickly.
• Experience shipping and being responsible for a production system that broke.
• Ability to take problems from concept to shipped product with minimal structure.
• Willingness to work across backend, infrastructure, and tooling as problems require.
• Ability to make reasonable decisions with incomplete information and adapt as more information becomes available.
• Fast gap-closing ability; preference for learning missing skills over avoiding them.
• Experience using AI tools in engineering workflows.

Nice To Haves

• Experience with distributed systems, queues, storage, or observability.
• Experience with systems that interact with numerous third-party APIs.
• Experience with LLM inference or agent architectures.
• Experience building and operating systems at scale, including personal side projects.

Responsibilities

• Work on routing and capacity, ensuring model calls meet latency, cost, and quality targets in real-time, and managing the underlying scheduling, reservations, and forecasting.
• Contribute to metering and billing correctness, dealing with multiple charging layers, hold and capture semantics, and reconciling a large volume of consumption events.
• Enhance reliability at sustained scale, focusing on observability, incident practice, and architecture to proactively manage growth.
• Develop the control plane, including credential-scoped permissions, workflow state limits, and gateway hardening.
• Improve agent execution, including sandboxed runs, state management, memory, tools, and external integrations.

What This Job Offers

• Job Type: Full-time
• Career Level: Entry Level
• Education Level: No Education Listed