SDE III

Responsibilities:

• Build and own production-grade backend services handling real user traffic.
• Design and build LLM agents involving LLMs, tool calling, and multi-step workflows.
• Own agents end-to-end from architecture and prompts to tools, evaluation, deployment, and post-production improvements.
• Design agent architecture covering tool orchestration, context/memory, guardrails, and decision-making on when to use an agent vs. traditional code.
• Build and maintain evaluation frameworks to measure agent quality and reliability.
• Work with golden datasets, regression suites, LLM-as-judge, human review, trajectory evaluations, and other evaluation approaches.
• Implement eval-gated releases, CI checks, shadow/canary deployments, and rollback mechanisms.
• Monitor production agents through tracing, drift monitoring, retries, timeouts, latency, and cost controls.
• Optimize model selection based on accuracy, cost, and latency, including experience migrating between models.
• Contribute to engineering excellence, architecture decisions, code quality, prioritization, and mentorship.
• Take ownership of production issues and drive fixes after launch.

Requirements:

• 6+ years of experience, primarily owning production backend services with real traffic.
• Strong backend engineering experience; language/technology stack is flexible.
• At least 1+ year of hands-on experience building LLM agents in production with real users.
• Experience with LLM + tool calling + multi-step agent workflows.
• Must have owned at least one agent end-to-end: design, prompts, tools, evaluations, deployment, and post-launch fixes.