关于本岗位：
我们在构建 AI Agent 中台，探索大模型能力如何被更可靠地编排、评测与工程化。你将和技术团队一起，在 Agent 架构、自动化评测和前沿技术预研中解决真实问题。
你将参与：
· 参与 AI Agent 中台的架构设计与核心功能开发，产出可运行、可迭代的工程实现；
· 协助构建 Agent 中台的自动化评测框架，对不同大模型的 Agent 能力进行深度评估，沉淀可复用的评测方法；
· 跟进业界最新的 Agent 论文与框架，参与内部技术预研，输出有落地价值的技术调研与原型验证。
我们希望你具备：
· 本科及以上学历；
· 掌握 Python/Java/C++ 等编程语言开发，具备扎实的算法与数据结构基础；
· 具备 Agent 系统开发经验，掌握 LangGraph/ADK/AgentScope 等框架及其实现原理；
· 了解 Claude Code/OpenClaw 等前沿 Agent 的实现机制，能设计高可用、高扩展性的大模型/Agent 工程架构；
· 掌握 Prompt 工程、RAG、Tool Calling、Harness Engineering、Skills 等 AI 应用关键技术；
· 了解机器学习、深度学习框架（PyTorch/TensorFlow）；
· 对 AI 技术充满热情，熟练使用各类 AI 提效工具，具备快速学习能力，持续关注 AIGC 前沿研究。
加分项：
· 计算机科学、软件工程、人工智能等相关专业背景；
· 有 Transformer 等模型实战经验；
· 在 NeurIPS、ICML、ICLR、ACL 等相关会议发表过论文。
What You’ll Be Part Of:
We’re building an AI agent platform and exploring how to make large-model capabilities more reliably orchestrated, evaluated, and production-ready. You’ll work alongside engineers on agent architecture, automated evaluation, and early-stage research, solving real problems rather than running toy demos. This internship is based in Xiamen.
How You’ll Contribute:
· Build and iterate on core components of the AI agent platform, turning designs into runnable, maintainable engineering implementations;
· Help design an automated evaluation framework for agent capabilities, test different large models against consistent benchmarks, and turn findings into reusable evaluation methods;
· Track and prototype emerging agent papers and frameworks, then translate what matters into internal technical surveys and proof-of-concept validation.
What We Hope You Have:
· Currently pursuing or already holding a Bachelor’s degree or above;
· Proficient in Python, Java, or C++, with solid foundations in algorithms and data structures;
· Hands-on experience building agent systems and familiarity with LangGraph, ADK, AgentScope, or similar frameworks, including how they work under the hood;
· Understanding of how Claude Code, OpenClaw, or similar leading agent implementations are built, and the ability to design high-availability, scalable LLM/agent architectures;
· Familiarity with prompt engineering, RAG, tool calling, harness engineering, and skills;
· Working knowledge of machine learning and deep learning frameworks such as PyTorch or TensorFlow;
· Genuine curiosity about AIGC research and AI-powered workflows, with a habit of using AI tools to speed up your own work.
Preferred Qualifications:
· Background in computer science, software engineering, artificial intelligence, or a related field;
· Hands-on experience with Transformer-based models;
· Publications at conferences such as NeurIPS, ICML, ICLR, or ACL.