We are hiring a Senior / Staff-level Agent Researcher or Engineer to work on evaluation-driven, system-oriented Agent and Multi-Agent applications in complex real-world scenarios.

What You Will Do

• Build and evolve evaluation-driven Agent systems, using evaluation as the primary mechanism to guide Agent behavior, iteration, and long-term improvement
• Design Agent systems that are observable, maintainable, and optimizable under real-world constraints
• Develop scalable and automated Agent evaluation infrastructure that can be reused across projects
• Continuously optimize Agent architectures, model choices, context, memory, and tool usage with clear performance–cost trade-offs
• Rapidly prototype Agent systems and demos using AI-native development practices (research × engineering × product validation)

What We Look For

• Senior / Staff-level experience with Agent, LLM, or complex AI systems
• Strong ability to track and internalize advances in Agents, LLM / MLLM, multimodal models, multi-agent systems, and context engineering, and translate them into practical solutions
• Solid foundations in AI / ML and systems, with a deep understanding of models, inference, training, context, evaluation, and system architecture (beyond API-level usage)
• Experience with Agent evaluation frameworks, automated experimentation, or A/B testing
• A research-oriented mindset combined with strong engineering execution
• Fluent Chinese communication skills enabling close collaboration with China-based teams

Nice to Have

• Publications, technical blogs, or open-source contributions
• Experience with automotive, chip, or intelligent cockpit systems (a plus, not a requirement)

Keywords

Evaluation-driven Agent · Systems Thinking · Research × Engineering · AI-Native

We are less interested in whether an Agent can simply run,and more focused on how well it performs, why it behaves as it does, and how it can continue to improve over time.