Required Graduation Period : 2026.9.1 - 2027.8.31

Intern Position

AI Agent Engineer

Position Overview

If you're not satisfied with simply calling APIs, and instead want to build intelligent Agents with real “thinking” capabilities, this is the perfect stage for you to play. Join our IT team to play a key role in the AI-native upgrade of the company's Agent platform, becoming a driving force in the democratization of technology. You won't just do coding — you'll build autonomous decision-making engines that enable complex business processes to evolve under AI-driven automation.

Key Responsibilities

• Deeply participate in the product research and development of an agent platform; explore how Agent reasoning chains and planning/execution design patterns can be implemented within the enterprise.
• For specific business scenarios, use mainstream Agent frameworks (such as Dify and LangGraph) to design multi-step task orchestration and workflows, enabling logical reasoning and complex task execution.
• Explore multi-agent collaboration frameworks, analyze complex business problems, and rapidly bring cutting-edge Agent capabilities into products.
• Explore and build robust evaluation systems (benchmark creation, automated evaluation, interactive capability evaluation) to continuously improve effectiveness, safety, and robustness.
• Technology iteration: stay current with industry advances (e.g., LangGraph, CrewAI, AutoGen), participate in AI-driven system upgrades, and complete the full cycle from prototyping and validation to production rollout and iterative improvement.
  
  

Qualifications & Skills

• Relevant background in computer science, artificial intelligence, software engineering, mathematics, or related fields; solid foundations in data structures and algorithms; strong understanding of operating systems and system design.
• Familiar with large language model (LLM) mechanisms (e.g., attention mechanisms, context window management); knowledgeable about optimization techniques such as Prompt Engineering, Chain of Thought (CoT), and RAG.
• Proficient in at least one backend programming language (Python/Java preferred); strong data structures and algorithms fundamentals.
• Hands-on experience with at least one mainstream Agent framework (LangGraph / LlamaIndex / CrewAI / Dify, etc.); understands how to call LLM APIs (e.g., OpenAI, Gemini, Qwen, DeepSeek, etc.).
• Familiar with core Agent technology stacks, such as:
• Agent patterns: ReAct, Plan-and-Execute, Multi-Agent collaboration
• RAG: vector databases, embedding models, retrieval strategy optimization
• Tool use: function calling, MCP protocol, dynamic tool orchestration
• Memory: short-term/long-term memory management, context compression
• Observability & evaluation: call-chain tracing, automated evaluation, benchmark building
• Safety: hallucination detection, prompt injection protection
• Strong ability to learn and explore quickly; excellent engineering and problem-solving skills; able to rapidly prototype, validate, and iterate.
  
  

Prefeed Qualifications (Nice To Have)

• You practice AI-native development and are skilled at collaborating efficiently with AI tools using natural language.
• You have unique insights into the Agent Development Lifecycle (ADLC) and have mature AI-assisted development workflow practices.
• Contributions to open-source projects, technical blogs, or publications in related fields.
• Passionate about thinking through and defining task boundaries, with a focus on continuous optimization of underlying architecture.
  
  

We look forward to working with passionate and innovative people like you to shape the future of AI Agents together!

Work Location:

Guangzhou

Interview City

Guangzhou/Online

※ You could apply one or two requisitions and identify it as “First Choice” or “Second Choice”. Once you submit your application, your choices will be finalized and cannot be changed.