Senior Backend Engineer (AI Agent)

Role overview Build backend AI agents for conversational workflows across the business commerce lifecycle, from product discovery and quoting to order handling and post-sale support. You will shape agent orchestration and tool use, while owning the evaluation practices that help establish whether an agent is reliable enough to release. The work spans early prototypes and production-ready systems.

Responsibilities - Design and iterate on agents in Go or Python, including orchestration, sub-agent delegation, intent handling, and safe execution of actions. - Build and maintain evaluation datasets, scorers, and regression suites; use results to improve measures such as product discovery and quote accuracy. - Treat prompts and model configurations as versioned, tested artifacts, and refine them based on evaluation and production results. - Integrate agents with MCP servers and backend tools, defining tool-call contracts, approval steps for write actions, and safeguards against misuse and prompt injection. - Add tracing, metrics, and cost monitoring, then use operational signals to guide improvements. - Collaborate with product and frontend partners, validate ideas with prototypes, guide architecture decisions, and mentor engineers in agent development.

Requirements - 5+ years of backend or full-stack software development experience, including hands-on work on agentic AI or LLM applications. - Strong programming ability in a major language such as Go, Python, Ruby, or PHP; a bachelor’s degree in a related technical field or equivalent experience is preferred. - Experience using LLMs in production and applying evaluation datasets, metrics, and results to improve systems. - Understanding of prompt engineering, tool or function calling, and agent orchestration patterns. - Experience designing APIs such as gRPC, GraphQL, or REST, integrating data across systems, and building MCP servers or other LLM tool integrations. - Strong problem-solving, communication, and collaboration skills, with a focus on execution in exploratory work.

Nice to have Experience with agent frameworks such as Google ADK or LangChain/LangGraph, and evaluation platforms such as Braintrust or similar tools.