Senior Performance Engineer

Atlassian•Mountain View, CA

2d•$149,400 - $235,000•Remote

About The Position

Atlassian is seeking a Senior Performance Engineer to join their Agentic AI Integration Products team. This role involves working on industry-leading products like the Rovo MCP Server and Agent to Agent integrations, leveraging cutting-edge AI capabilities such as Claude Skills and ChatGPT/Claude Apps. The ideal candidate is passionate about exploring and learning AI transformative technologies, quickly pivoting from prototyping to building highly-scalable enterprise-grade AI products used by thousands of developers and enterprise users. The role requires strong acumen in ML performance, quality, and systems, with experience in tuning servers for latency, reliability, token efficiency, and tool-selection quality. This includes expertise in dynamic tool discovery, context and response optimization, observability, automated evaluations, and semantic retrieval using embeddings, vector search, hybrid ranking, and reranking.

Requirements

7+ years of software engineering experience building and operating enterprise systems, APIs, or cloud-native products.
Strong proficiency in TypeScript/JavaScript and modern front-end development with React; experience building accessible, responsive, and performant web applications.
Hands-on experience designing or integrating MCP servers, tools, agent-facing APIs, or related context and agent frameworks.
Demonstrated ability to apply distributed-systems principles, including concurrency, connection pooling, caching, retries, timeouts, backpressure, autoscaling, and load shedding.
Experience measuring and improving latency, throughput, saturation, error rates, availability, token consumption, and end-to-end task cost.
Practical experience with AI/ML evaluation and quality engineering, including benchmark design, groundedness, relevance, safety, hallucination detection, bias analysis, monitoring, and regression prevention.
Knowledge of semantic search and retrieval systems, including embeddings, vector databases or indexes, hybrid retrieval, ranking, reranking, and permission-aware filtering.
Experience integrating GraphQL, REST, JSON Schema, streaming APIs, SDKs, and event-driven systems into reliable product experiences.
Proficiency in at least one additional systems or back-end language such as Python or Go, with strong testing and API design practices.
Proven ability to lead cross-functional engineering initiatives, communicate clearly with technical and non-technical partners, and mentor other engineers.

Nice To Haves

Familiarity with MCP architecture, A2A specification, and agent collaboration frameworks (e.g., MCP servers, UI clients, and adapters).
Experience with observability, vector databases, and secure model-to-model communication.
Background in enterprise integration patterns, API governance, or developer experience platforms.
Contributions to open-source AI frameworks or standards development initiatives.

Responsibilities

Design, build, and evolve MCP servers, tools, and agent-facing APIs with concise schemas, predictable errors, safe mutations, and clear outcome-oriented contracts.
Develop accessible, responsive, and performant React and TypeScript experiences that make agent capabilities, MCP tools, and A2A interactions easy to discover, configure, and use.
Build reusable components, design-system patterns, and frontend architecture that support consistent, scalable user experiences across AI-powered products.
Integrate GraphQL and REST APIs, SDKs, streaming responses, and real-time data into reliable, user-friendly AI workflows.
Optimize token and context efficiency through dynamic tool discovery, lazy loading, bounded responses, pagination, selective field retrieval, caching, and reduced tool-call loops.
Improve end-to-end performance and reliability across front-end clients, gateways, MCP servers, search services, and downstream product systems through observability, tracing, SLOs, and production diagnostics.
Build semantic retrieval capabilities using embeddings, chunking, vector indexes, hybrid search, metadata and permission filters, ranking, reranking, and freshness strategies.
Define and operate AI/ML quality programs with JTBD-based evaluations, benchmark datasets, groundedness and relevance metrics, hallucination and bias detection, safety testing, and human feedback.
Integrate automated evaluations into CI/CD and release gates to detect regressions across model, prompt, tool, and retrieval changes.
Implement enterprise security and partner cross-functionally to deliver maintainable, well-tested AI integrations, including OAuth 2.1, tenant isolation, audit logging, prompt-injection defenses, and confirmation flows for high-impact actions.