Job description
---
1. Full-stack delivery of business systems: Independently complete front-end (React/Next.js/Vue) and back-end (Node.js/Python/Java/Go) development for B-end/C-end web applications, management backends, and data platforms, covering database design, API development, performance optimization, containerized deployment, and online operations.
2. Native development of AI tools: Use AI programming tools such as Cursor, Claude Code, GitHub Copilot, v0 as daily development infrastructure, leveraging AI assistance for code generation, architecture design, unit testing, code review, and documentation writing to ensure a dual enhancement of development efficiency and code quality.
3. Design and implementation of AI workflows: Design multi-step AI workflows for business scenarios, including but not limited to: RAG retrieval-augmented systems (document parsing → embedding → vector retrieval → generation), agent task orchestration (tool invocation / multi-round decision-making / human-machine collaboration), intelligent content generation, and automated data processing pipelines.
4. Large model integration and prompt engineering: Interface with mainstream large model APIs such as OpenAI / Claude / Tongyi Qianwen / DeepSeek to achieve streaming output (SSE), function calling, and structured data extraction; write high-quality business-level prompts to ensure output stability and boundary safety.
5. Data and infrastructure architecture: Responsible for the architecture design and optimization of PostgreSQL / MySQL, Redis, and message queues; integrate vector databases (pgvector / Pinecone / Qdrant) to support AI retrieval capabilities; use Docker and CI/CD to complete automated testing and continuous delivery.
6. Observability and optimization of workflows: Establish an evaluation and monitoring system for AI workflows (output accuracy, token cost, response latency, exception degradation), implement safety guardrails and human-in-the-loop mechanisms to ensure the reliability of AI functions in production environments.
Skill requirements:
1. Full-stack development capability: Over 5 years of full-stack development experience, proficient in at least one front-end framework (React / Next.js / Vue) and one back-end language (Node.js / Python / Java / Go), able to independently complete the full functional loop from database design to front-end interaction.
2. Proficiency in AI tools: Regularly use AI programming tools such as Cursor / Claude Code / GitHub Copilot, familiar with best practices for AI-assisted programming (context management, code review, prompt-driven refactoring), able to quickly convert AI-generated code into production code that meets team standards.
3. Experience with AI workflows: Practical experience in implementing AI workflows in real projects, understanding RAG architecture, agent orchestration logic, and multi-step LLM pipeline design; hands-on experience with LLM API integration, streaming, and structured output.
4. Prompt engineering: Solid prompt engineering skills, able to design stable and reusable prompt templates for business scenarios, handle edge cases and adversarial inputs, ensuring AI output meets business expectations.
5. Engineering literacy: Solid coding standards, Git workflows, unit testing, and API design capabilities; familiar with SQL optimization and index design; understanding of Docker containerization and CI/CD processes; basic awareness of security protections (XSS / SQL injection / access control).
【Bonus points】
Familiarity with AI application development frameworks such as Dify, LangChain, LlamaIndex, LangGraph, able to quickly build complex AI workflows.
Experience in integration and retrieval optimization of vector databases (pgvector / Pinecone / Qdrant / Weaviate).
Practical experience in local model deployment (Ollama / vLLM) or model fine-tuning (LoRA / QLoRA).
Understanding of MCP (Model Context Protocol) or A2A (Agent-to-Agent) protocols, with cross-system AI integration capabilities.
Familiarity with automation orchestration platforms such as n8n, Make, Airflow, able to integrate low-code workflows with code-level systems.
Experience in building AI workflow evaluation frameworks (Golden Set, regression testing, quality drift monitoring).
---