Senior Software Engineer (AI / Agent) - 資深軟體工程師

What you’ll do:

• Architect durable agent workflows (using Mastra, LangGraph, or Temporal) and lead harness engineering, including designing the context, tools, and environments for reliable execution.
• Build shared AI capabilities, collaborating with Product and Design to stream real-time agent state, tool calls, and human-in-the-loop (HIL) approvals to the UI via AG-UI.
• Optimize for production scale and safety, managing model gateways, routing, token economics, latency, and guardrails against prompt injection.
• Make evaluation core to development by using tracing tools (like Langfuse) to measure agent behavior against real customer outcomes.
• Explore and prototype internal coding agents that can autonomously test and develop features to accelerate our engineering team.

Must-Have Qualifications:

• Backend Engineering: 5+ years of software engineering experience focusing on Python and/or TypeScript, building distributed backend systems or APIs.
• Agent Architecture & Durable Execution: Deep, hands-on knowledge of LLM orchestration, context memory, tool calling, and building long-running workflows that handle retries and external dependencies.
• Product Sense: Experience building interactive agent experiences that stream real-time state, tool execution, and HIL prompts to a frontend UI.
• Production AI: Practical understanding of managing AI safety (e.g., prompt injection) alongside system performance (cost, latency, token limits).
• Navigating Ambiguity: Comfortable working in a fast-moving field, prototyping new research, and finding a path forward without a documented playbook.

Preferred Qualifications:

• Hands-on experience with Mastra, Temporal, or LangGraph for orchestrating complex, multi-step agents.
• Experience with harness engineering and Model Context Protocol (MCP) for securely managing agent environments, tool permissions, and authentication.
• Experience with AG-UI (or similar state-streaming protocols) to move agent state, approvals, and errors to the product frontend.
• Familiarity with LLM observability (Langfuse) and AI infrastructure (model gateways, prompt caching, multi-provider routing).
• Built or used autonomous coding agents in day-to-day engineering (e.g., harnesses that let agents implement work, test it, and generate evidence for review).

You will be a good fit if you:

• Enjoy freedom and take responsibility in a hybrid environment.
• Are excellent at communicating clearly.
• Know how to effectively and efficiently collaborate with people around the world (cross timezones).
• Can manage multiple projects concurrently and prioritize effectively.
• Flexible attitude and willingness to frequently move between different teams, software architectures and priorities.
• Desire to collaborate with our other product teams to think strategically about how to solve problems.

Updated: 2026/10/08