SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Sapiom is building the operating system for agentic AI systems. The company has raised $50M (Series A led by Dragonfly in August 2026) and is backed by Accel, Menlo Ventures, and Anthropic. The founder previously led payments engineering at Shopify and built an autonomous consumer agent company.
The Agent Platform is the infrastructure layer that makes agentic products production-ready. While models provide the intelligence, Sapiom provides the operational spine: compute and sandboxes, memory, identity, spend controls, storage, queues, and monitoring—unified as a single platform rather than a framework to assemble.
In this full-stack senior role, you would own one system end-to-end, responsible for both backend infrastructure and the customer-facing interfaces that give teams visibility and control. The platform currently processes 270M+ consumption events and has absorbed 10–20x growth compounding week-over-week.
Key areas of focus include:
**Routing and Capacity**: Place every model call against latency, cost, and quality targets in real time. Build durable scheduling, reservations, and forecasting to keep decisions sound as demand shifts. One customer achieved a 75% reduction in inference costs through this system.
**Metering and Billing Correctness**: Manage two charging layers with hold-and-capture semantics across 270M+ consumption events. Ensure numbers reconcile in both the customer dashboard and the ledger. Correctness is a core product differentiator.
**Reliability at Scale**: Design for sustained growth and open questions around what reliability means at this scale. Improve observability, incident practice, and architecture to stay ahead of growth rather than behind it.
**Control Plane**: Implement credential-scoped permissions, workflow state limits, gateway hardening, and audit surfaces. Enable customers to safely delegate specific authority to agents rather than blanket access.
**Spend Controls**: Build budgets enforced before money moves and reporting that makes agent spending explicable. This often determines whether enterprise pilots convert to deployments.
You would carry the pager for your system and work closely with a world-class infrastructure and payments-focused team. The engineering team uses the internal title "Member of Technical Staff."
**Requirements**
- 4–7 years of experience
- Owned a production system end-to-end and carried the pager for it
- Experience in domains where being wrong is expensive: billing, payments, ledgers, or systems that must reconcile against external numbers
- Designed for failure upfront (timeouts, retries, idempotency, backpressure) and can articulate a case where your design failed and what you changed
- Debugged novel production issues under time pressure and documented the findings
- Led multi-month projects from problem statement (not spec) and shipped work that other engineers built upon
- Build interfaces on top of your own systems rather than handing that off
- Use AI tools in your own engineering workflow
**Nice to have**: metering, billing, or usage-based pricing at scale; payments or financial systems; API gateways, proxies, or traffic management; LLM inference, model routing, or capacity planning; queues, storage, or observability infrastructure; systems integrating many third-party APIs; early-stage experience setting architecture rather than inheriting it.
