Principal Engineer, Backend Engineering

Atlassian•San Francisco, CA

52d•$174,051 - $273,775

About The Position

Atlassian's internal AI usage is growing rapidly across engineering, driven by adoption of tools like Codex, Cursor, Claude Code, and Rovo CLI. This growth is a signal of value, and we need a Principal Engineer to build the infrastructure that makes it sustainable. You will own the technical vision and architecture for the DevInfra AI Front Proxy, a cloud-hosted service that sits at the developer-workflow boundary and governs every internal model call against cost, trust, context, and attribution constraints. The proxy spans three integration surfaces: managed local configuration pushed via MDM, the cloud proxy service itself, and a browser extension for web AI interfaces. It is the single layer through which all internal AI traffic flows for cost enforcement, security policy integration, model routing, semantic caching, and telemetry. This is a high-impact, high-ambiguity role. You will partner with teams across developer infrastructure, AI platform, security and compliance, and developer tooling to define and ship the platform that makes exponential AI usage growth safe, measurable, and cost-efficient at Atlassian scale.

Requirements

High-impact, high-ambiguity role requiring ownership of technical vision and architecture.
Partnering with multiple teams including developer infrastructure, AI platform, security and compliance, and developer tooling.
Defining and shipping a platform for safe, measurable, and cost-efficient AI usage growth.
Experience with cloud-hosted services and developer workflow integration.
Understanding of cost enforcement, trust, context, and attribution constraints for AI model calls.
Experience with managed local configuration, cloud services, and browser extensions.
Knowledge of cost enforcement, security policy integration, model routing, semantic caching, and telemetry.
Ability to design and deliver highly available (99.95%+) and low-latency (sub-second) services.
Expertise in cost-efficiency strategies including quota enforcement, semantic caching, context reuse, model routing, and cost attribution.
Experience defining security integration surfaces for policy enforcement across AI tools.
Skills in context integration, including connecting with context infrastructure, repository harness, and enabling skill/tool registry sharing.
Experience with cloud infrastructure integration for model execution, output token optimization, intelligent model routing, and context compression.
Strong partnership and cross-functional coordination skills.
Mentorship and team leadership experience.
Experience establishing and owning telemetry and observability stacks.

Responsibilities

Set the technical direction for the AI proxy platform across three pillars: cost efficiency and ROI insights, secure AI usage, and AI context integration.
Design and deliver the proxy service architecture targeting 99.95% or higher availability, while maintaining sub-second latency to preserve developer experience across all routed AI traffic.
Own the cost-efficiency stack: integration with internal usage quota enforcement systems with near real-time cap enforcement, semantic caching and context reuse across tools, model routing and downgrade policies, and workflow-level cost attribution.
Define the security integration surface so that security, compliance, and supply chain governance teams can enforce policies at machine speed across all internal AI tools through shared hooks, rather than building bespoke runtimes per tool.
Drive the context integration layer: connecting with context infrastructure and repository harness, enabling skills and tool registry sharing across AI tools, progressive disclosure of context, and agent quality benchmarking across tools and repositories.
Lead technical decisions on cloud infrastructure integration for model execution, output token optimization strategies, intelligent model routing, and context compression tooling rollout.
Partner with AI platform, security, compliance, and developer tooling teams to maintain alignment across a multi-team initiative with significant cross-functional coordination requirements.
Mentor and uplift a team of engineers across multiple delivery streams covering reliability, usage governance, and proxy core, while reserving sufficient capacity for system reliability alongside feature delivery.
Establish and own the telemetry and observability stack: dashboards for agent and CI traffic, cache hit rates, workflow effectiveness, cost per successful workflow, and ROI attribution by tool, team, and user.