Staff Software Engineer, AI Gateway

About Docker
Docker has been one of the most loved brands in developer tooling, trusted by more than 20 million monthly users and over 20 billion container image pulls. From solo founders to the world's largest companies, developers rely on Docker to build, share, and run their applications across our suite of products including Docker Desktop, Docker Hub, and Docker Scout. We are a globally distributed, remote-first team building the tools that define how software gets built and delivered. As AI agents redefine software development, Docker is at the center of that shift, providing the sandboxed environments, verified images, and secure infrastructure that make autonomous workflows trustworthy by default.
_______________________________________________________________________

Docker AI Gateway is our answer to the complexity of taking AI agents from prototype to production. It’s a powerful, intelligent, and secure control point that eliminates the toil of model orchestration, tool management, observability, and governance - so developers can focus on building incredible AI agents, not gluing together infrastructure.
The Gateway sits at the center of modern AI applications, offering:
A model and tool routing layer with built-in security and cost optimization

A familiar OpenAI-compatible interface and MCP server

Unified observability and policy enforcement

Auto-RAG, tool injection, session summarization, and more

We’re just getting started and we need exceptional engineers to help us build the backbone of the future of agent-based development.

What You’ll Do
Set technical direction for core systems powering the AI Gateway 0 the model router, MCP gateway, and control plane - and make sure architectural decisions hold up as usage scales by 10x

Consistently identify gaps in our infrastructure, process, or tooling before they become blockers, and drive the fix from idea to production

Take a significant, hands-on role in defining, delivering, and implementing the Gateway's technical strategy, in close partnership with engineering leadership

Own the hardest, highest-ambiguity problems in the system: dynamic model selection, auto-failover, cost-based routing, and policy enforcement that has to work end-to-end

Act as the go-to expert for secure credential storage, session summarization, caching, and rate limiting - the capabilities everything else depends on

Build the infrastructure that lets us, and our customers, evaluate, replay, and backtest agent and LLM behavior with confidence

Mentor senior and mid-level engineers, and raise the technical bar for the team through code, design review, and example

Partner with product, design, and leadership to shape not just what the control plane UI does, but where the product goes next

Key Problems You’ll Help Solve
Define the unified abstraction layer across diverse model and tool providers (OpenAI, Anthropic, Google, AWS Bedrock) that the rest of the org builds on

Architect secure, scalable identity and credential vaulting for tool and model access, trusted enough to run at enterprise scale

Build the infrastructure for real-time and historical analytics of AI agent behavior that leadership and customers both rely on

Guarantee that policy enforcement and logging work end-to-end - from prompt to tool to response - with no gaps an auditor (or attacker) could find

Set the bar for developer experience through intuitive APIs and first-class observability

Who You Are
Required Qualifications
8+ years of backend engineering experience, with a track record of owning production-grade systems at significant scale

Deep specialization in distributed systems, cloud-native infrastructure, and API design - recognized by peers and cross-functional partners as a go-to expert, not just experienced

Experience building secure, high-throughput services (gateways, proxies, load balancers, policy engines), and a history of making judgment calls other engineers rely on