For more details, please visit the job posting link:https://m.zhaopin.com/xiaoyuan/company/detail?redirect=mxiaoyuan&comid=KA0139131344P90000003000&productId=-1&channelId=-1

I. Job Responsibilities<br>Tackling Core Systems: Use AI tools (Cursor/Claude, etc.) to directly code and verify architectural ideas, allowing AI to assist in engineering implementation. You are responsible for defining the core logic and performance limits of the value definition, value measurement, and operation of intelligent agents in the Internet of Things, without acting as a mere mouthpiece for requirements.<br>Designing Distributed Workflows: Decompose intelligent agent value mining scenarios into an autonomous workflow of data collection - value extraction - operational closed loop - verification iteration, ensuring the system runs stably like a high-efficiency engine.<br>Implementing Value Operations: Design high-concurrency, highly available value mining and operation systems for the global intelligent agent ecosystem, balancing system performance and business implementation efficiency.<br>Driving AI-Native Iteration: Based on system operation data, use natural language to guide AI to adjust architecture, optimize code, and refactor services, achieving daily iterative releases.<br>Defining New Engineering Paradigms: Explore engineering collaboration methods in the AI era, rejecting repetitive CRUD operations, allowing AI to automatically complete coding, testing, and deployment, delivering system results using dialogue + real-time architecture diagrams.<br>II. Hard Thresholds<br>AI Native Proof<br>Using AI The tool can independently deliver at least one online distributed system function, demonstrating the code/architecture diagrams generated by AI and the core decisions you made.<br />Having personally written complex distributed systems/agent workflows, optimized high-concurrency scenarios, and clearly understanding the capability boundaries and optimization methods of distributed systems.<br />Not "requiring the team to develop AI functions," but rather "getting the system running with AI first before handing it over to the team for implementation."<br />Possessing distributed system development experience, proficient in core distributed technologies such as microservices, caching, and message queues.<br />Able to use AI tools proficiently for programming, leveraging AI to improve coding, debugging, and architecture design efficiency.<br />Possessing system architecture design capabilities, able to independently complete the architecture planning and implementation of large-scale distributed systems.<br />Having served as a development team leader, proficient in software engineering, and having practical experience in leading teams in collaborative development.<br />III. The Qualities We Seek<br />Soft Qualities (Geek Engineer Temperament)<br />Engineering Taste: Born With Taste, naturally sensitive to clean code and elegant architecture, restless at the sight of bad code, and willing to discuss 100+ things with AI for 0.1ms performance optimization. **Technical Sense:** Upon learning of the latest large-scale models/distributed framework releases, the first thought is "how to use it in my system." Able to discuss the latest papers/infrastructure/technology stacks with the algorithm team. **Hands-on Obsession:** Disdains lengthy review meetings, preferring runnable distributed demos; hates repetitive work, automating tasks for AI; rejects old-school architects' empty talk, adhering only to PMF (Product-Market Fit) first principles; makes decisive, unhurried decisions, and goes live to analyze data. **Combat Experience:** Has submitted PRs related to distributed systems in open-source communities, gathered user feedback on Discord, showcased AI engineering products on Product Hunt, or built internal and external distributed systems using agents. **Top-Tier Hunger:** Only sees the world's top AI engineering standards; feels that if a system doesn't impress Silicon Valley peers, all efforts have been wasted. **Cultural Compatibility:** Has read "Clean Code" and can discuss architectural aesthetics; accepts the "if it's not good enough, refactor" scaling principle. Iterative brute force, maintaining ignorance and a geeky spirit.<br>IV. What we offer:<br>Zero bureaucratic decision-making power: Your distributed demo goes directly to the technical lead for testing. If it works well, it goes live. There are no three-tier reporting systems or requirement review meetings. System architecture decisions are in your hands, and you allocate technical resources.<br>Global influence: Your systems will be used by real intelligent agents for value mining and global operations. Your work may appear in the acknowledgments of top industry reports.<br>Hardcore and mysterious colleagues: The colleagues sitting next to you may be former top distributed systems engineers or former AI agent architects who have experienced the kind of technical challenges you only see in movies. They will point out your bugs and criticize your architecture, but three months with them is equivalent to three years of industry experience.