Software Engineer, Full Stack (Agent Platform)

About the role

Sapiom is building an end‑to‑end Agent Platform that unifies compute, memory, identity, spend controls, storage, and monitoring for autonomous agents. As a senior full‑stack engineer you will own a critical piece of this platform, delivering reliable, cost‑effective, and observable services that power production agents.

Key responsibilities

• Routing and capacity: place every model call against latency, cost, and quality targets in real time, and implement durable scheduling, reservations, and forecasting.
• Metering and billing correctness: manage two charging layers, hold and capture semantics, and reconcile 270M+ consumption events with high accuracy.
• Reliability at sustained scale: handle 10–20× growth, define observability, incident practices, and architectural changes for large‑scale reliability.
• Control plane: design credential‑scoped permissions, workflow limits, gateway hardening, and audit surfaces for customer‑configured authority.
• Spend controls: enforce budgets before money moves and provide clear post‑spend reporting for enterprise deployments.

Required profile

• Owned a production system end‑to‑end and carried the pager for it.
• Worked in environments where errors are costly (billing, payments, ledgers).
• Designed for failure up front (timeouts, retries, idempotency, backpressure) and can cite a real fix.
• Debugged unseen production issues under pressure and documented the incident.
• Led multi‑month projects from problem statement to shipped solution that others built upon.
• Built user‑facing interfaces on top of your own systems.
• Use AI tools in your engineering workflow.

Required skills

• Metering and usage‑based billing systems.
• API gateways, proxies, or traffic management.
• LLM inference, model routing, or capacity planning.
• Queues, storage, and observability infrastructure.