We're looking for a Senior Engineer to build Qase's MCP server and Qase Harness: the developer tools AI coding agents use to work with test management.

About Qase

Qase.io is a fast-growing Series A startup (headed towards Series B) building test management SaaS trusted by engineering teams at companies like Asana, Crypto.com, SUSE, SeatGeek and Wolt.

We're remote-first, with a fully distributed team across several countries. We're around 50 people, founded in 2019.

About the Role

AI coding agents now write code and tests faster than anyone can review them. They don't know what "safe to ship" means for a product: what's covered, which requirements exist, what failed before, which standards the team holds. Qase is building the quality layer agents work through to get that context and verify their work.

You'll own the two pieces that make this real:

• Qase MCP connects agents to Qase.
• Qase Harness ships as a plugin for Claude Code, Cursor and other agents, carrying Skills, Context and Standards that tell an agent what to do and how to do it well.

We build this as a product, not as internal tooling: it ships to customers, and its users are both engineers and the agents they run.

What You'll Do

• Own the Qase MCP server (https://www.qase.io/mcp-model-context-protocol/) end to end.
• Build Qase Harness: the plugin that packages Skills, Context and Standards for Claude Code, Cursor and other agents.
• Design tools agents use well: clear schemas, token-efficient responses, predictable errors, safe defaults and scoped permissions.
• Evolve our CLI (qasectl), API and reporters so agents can query Qase mid-loop, not just push results out.
• Build verification loops that check an agent's changes against coverage, requirements and past failures.
• Set up evals for tools and skills as CI gates, and measure what matters: tool-call success, latency, token cost and adoption.
• Harden agent-facing surfaces against prompt injection and over-broad actions, with an audit trail of what agents did.
• Run it like a product: docs, examples, versioning, changelogs, usage telemetry, and direct conversations with the customers who use it.

What We're Looking For

• 10+ years of software engineering, and still strongly hands-on.
• A track record of shipping developer-facing products (CLIs, SDKs, public APIs, IDE extensions or plugins) where developer experience was the point.
• Daily hands-on use of coding agents like Claude Code or Cursor, and a clear view of how agents pick tools and where they fail.
• Solid API design experience.
• Product sense: you decide what to build from how people actually use it, not just from a ticket.
• A quality mindset: you care how things get verified, not just how they get written.
• Strong written and spoken English.

Nice to Have

• Building and running an MCP server or other agent tooling in production, with a solid grasp of the MCP spec (tools, resources, prompts, transports, auth).
• Experience with test automation frameworks and reporters (Playwright, pytest, JUnit, Cypress) or test management tools.
• Experience with LLM eval tooling.
• Architecting CI/CD pipelines from scratch (build, test, release, gating, caching, environments).
• Agent observability, for example OpenTelemetry GenAI semantic conventions.
• Maintaining open-source developer tools or plugins.
• Platform engineering experience (internal platforms, shared tooling).
• Writing technical docs, tutorials or talks for developers.
• A drive to spread agentic practices across the company, helping other teams adopt shared skills, tools and ways of working with agents.

What Makes This Interesting

• The chance to define how AI agents work with test management, in a category that's still taking shape.
• A real product: what you build ships to customers and their agents, not just to our own team.
• Full ownership of MCP and Harness, with no layers between you and decisions.