Software Engineer, Agent Experience

About the role:

More and more developers don't read our docs or click through our dashboard. Their coding agent does. Claude Code, Cursor, Codex and others are now how a big share of engineers discover, integrate and debug AssemblyAI. We've spent years polishing the human experience. The agent experience (AX) is newer, growing fast, and almost untouched.

That's the job. You'll own how AI agents build with AssemblyAI: the skills, prompts, MCP server, code samples, docs and integrations an agent reads and runs when a developer says "add transcription to my app." Your north star is simple: an agent should go from zero to a working, production-grade AssemblyAI integration on the first try.

You'll work closely with senior engineers on our Product Engineering team, with real ownership from day one and room to grow into leading AX as the area scales.

What You’ll Do:

• Drive the agent experience. Treat coding agents as a first-class user of AssemblyAI, alongside the humans who use our dashboard and Playground.
• Build and maintain agent skills. Ship skills for our core products (Speech-to-Text, Real-time, Dictation, Voice Agents, TTS, LLM Gateway) that teach agents how to integrate us correctly.
• Build on our MCP server and agent integrations. Extend the AssemblyAI MCP server, help ship agent and MCP auth (OAuth and API keys), and make it easy to plug into Claude Code, Cursor, Codex and whatever comes next.
• Make our docs and code samples agent-ready. Own llms.txt, machine-readable docs, quickstarts and SDK examples so agents pick the right API, parameters and patterns the first time.
• Tune the prompts and instructions behind our product. Improve the system prompts, tool definitions and agent instructions that power our in-dashboard agent and onboarding flows.
• Measure it. Build evals and benchmarks that run real coding agents against real integration tasks. Track success rate, time to first working call, and activation from agent-driven signups.
• Help us scale the playbook. Start with one product and then roll what works across every product.
• Use AI tools as a core part of how you work, and help the rest of engineering build faster with agents.

What You’ll Need:

• 2+ years of professional software engineering experience, with solid full-stack fundamentals (Javascript and/or Python).
• Daily, hands-on use of coding agents such as Claude Code, Cursor or Codex, and real curiosity about what makes them succeed or fail.
• Some hands-on experience writing prompts, skills, tool definitions or agent instructions, at work or side projects, and an interest in measuring them with evals rather than vibes.
• Product sense for developer-facing work. You care whether the developer, or their agent, actually gets to a working integration.
• Clear technical writing. Much of this role is writing that agents and humans both read.
• Comfort with ambiguity and a bias to ship. AX is a new discipline, and you'll learn it and help define it alongside us.

Nice to Have:

• Working knowledge of MCP, LLM tool calling, and how agents read docs and APIs.
• You've built or contributed to an MCP server, agent skill library, or coding agent plugin.
• Developer relations, docs engineering, or SDK experience at an API-first company.
• Experience with Voice AI, speech-to-text, or LLM APIs.
• You've shipped something that agents use in the wild, and can show us the before and after.

The provided ranges are the expected salaries for candidates in the United States (USD) and Canada (CAD). Outside of those regions, there may be a change in the range which will be communicated to candidates throughout the interview process.

USD Salary Range: $160,000 - $210,000

CAD Salary Range: C$225,000 - C$285,000