Senior Full Stack Engineer - Agent Runtime & Platform

NinjaTech AI; Sydney; Hybrid; AUD 140,000 - 200,000; Full time

Job summary

NinjaTech AI is seeking a backend platform engineer in Sydney to own the agent sandbox runtime and related platform services. You will manage provisioning, isolation, scheduling, and cost per task at scale, while building the instrumentation and tools that keep the system observable and reliable.

The role emphasizes Python proficiency, AWS CDK and container orchestration expertise, with a pragmatic approach to design and implementation for scalable growth.

Qualifications

• 5+ years building and operating production backend systems and platforms.
• Strong AWS CDK and container orchestration understanding.
• Python strength with frontend basics for internal tooling.
• Bias for simplicity over architecture for its own sake.

Responsibilities

• Own the agent sandbox runtime: provisioning, isolation, scheduling, lifecycle and cost per task at scale.
• Build and own platform services — identity plumbing, secrets, metering, object storage, workflow orchestration, and the tool/connector abstraction layer.
• Drive the observability stack from instrumentation through ingestion to the dashboard people actually open during an incident.
• Harden infrastructure against the failure modes that surface to users as “the agent got stuck”, and keep the platform from becoming a bottleneck to growth.

Skills

Backend systems

AWS CDK

Container orchestration

Python

Frontend basics

Tools

AWS CDK

Job description

Sydney

• Full-time
• On-site/hybrid

Every agent task runs somewhere: a sandboxed container with a filesystem, a browser, tools and a time limit. This role owns that runtime and the platform services around it — the layer everything else in the product depends on.

What You\'ll Do

• Own the agent sandbox runtime: provisioning, isolation, scheduling, lifecycle and cost per task at scale.
• Build and own platform services — identity plumbing, secrets, metering, object storage, workflow orchestration, and the tool/connector abstraction layer.
• Drive the observability stack from instrumentation through ingestion to the dashboard people actually open during an incident.
• Harden infrastructure against the failure modes that surface to users as “the agent got stuck”, and keep the platform from becoming a bottleneck to growth.

What We\'re Looking For

• 5+ years building and operating production backend systems and platforms.
• Deep AWS (we are CDK-first) and container orchestration understanding below the level of configuration — schedulers, runtimes and isolation boundaries.
• Python strength, plus enough front end to build the internal tooling your platform needs rather than waiting for someone else.
• A bias for simplicity and pragmatism over architecture for its own sake.