QA Automation Engieer (MCP and AI Infrastructure)
Who recruits here
Up to three recruiters at Photon with their work email, so you can write to a person.
Available on Pro.
Description
Job Summary
We are seeking an experienced QA Automation Engineer to build and maintain automated testing solutions for Model Context Protocol (MCP) servers and AI agent integrations. The role focuses on functional testing, protocol compliance, performance testing, and reliability to ensure production-ready AI infrastructure.
Key Responsibilities
Develop automated MCP test harnesses using TypeScript or Python to validate tools, resources, prompts, and JSON-RPC error handling.
Implement functional, schema validation, and contract regression tests within CI/CD pipelines.
Design and execute load, stress, soak, and concurrency tests to evaluate MCP server performance and reliability.
Build evaluation pipelines to validate AI agent tool selection, multi-step workflows, and failure recovery.
Establish logging, monitoring, and alerting for server health, authentication, and rate limiting.
Collaborate with AI and engineering teams to improve test coverage and production reliability.
Required Skills and Qualifications
5+ years of experience in QA Automation, Software Testing, or SDET roles, particularly with backend APIs and distributed systems.
Hands-on experience testing MCP servers using MCP client SDKs or custom JSON-RPC test frameworks.
Strong programming skills in TypeScript/Node.js or Python.
Experience with testing frameworks such as Vitest, Jest, or Pytest.
Practical experience with performance testing tools such as k6, Locust, or JMeter.
Experience integrating automated tests into GitHub Actions, GitLab CI, or Jenkins.
Knowledge of Docker and CI/CD pipelines.
Strong analytical, troubleshooting, and communication skills.
Bachelor’s degree in Computer Science or equivalent.
Preferred Qualifications
Experience testing Streamable HTTP, SSE, WebSockets, and stateful sessions.
Familiarity with OAuth 2.1, API security, and sandboxing.
Experience with LLM agent evaluation, tool-calling benchmarks, or prompt regression testing.
Skills
Sign in to read and join the discussion.