大模型应用开发工程师LLM Application Engineer

大模型应用开发工程师 LLM Application Engineer
我们正在寻找一名大模型应用开发工程师，负责基于大语言模型的应用设计、结构化数据解析与优化工作。
We are seeking a highly skilled LLM Application Engineer to lead the design, structured data parsing, and optimization of applications powered by Large Language Models.

职位类型 / Job Type
远程兼职职位。工作时间灵活，首期项目周期预估 2 个月（表现优异者可拓展长期合作），薪资按阶段/里程碑进行支付。
This is a remote, part-time position with flexible hours. The initial project duration is estimated at 2 months (potential for long-term collaboration), with payments structured around project milestones.

岗位职责 / Responsibilities
1.集成和调用主流大语言模型（如 GPT、Claude、Gemini、Ollama 等），实现基于语义理解的高精度结构化信息抽取。
Integrate and invoke leading LLMs (e.g., GPT, Claude, Gemini, Ollama) to achieve high-precision structured information extraction based on semantic understanding.
2.设计并实现批量 HTML 或文档语义解析 Pipeline，并将其与底层数据抽取规则引擎深度结合。
Design and implement semantic parsing pipelines for batch HTML/document processing, ensuring seamless integration with data extraction rule engines.
3.持续优化大模型提示词工程（Prompt Engineering）模板，通过迭代不断提升输出结果的准确性。
Continuously optimize Prompt Engineering templates to improve the accuracy and robustness of model outputs through iterative testing.
4.负责结构化结果（JSON/CSV/数据库）的管理与存储，为后续数据分析与 API 接口输出提供坚实支撑。
Manage and store structured results (JSON, CSV, or databases) to support downstream data analysis and API delivery.
5.构建并行解析架构和完善的错误重试机制，在高通量运行环境下保障系统的高可用性与稳定性。
Build parallel processing architectures and robust error-retry mechanisms to ensure system stability and high throughput during large-scale operations.
6.编写并维护可复用的技术脚本与开发文档，支持在多种业务场景下的快速部署与模块复用。
Develop and maintain reusable technical scripts and documentation to facilitate rapid deployment and module reuse across various scenarios.
7.独立闭环完成从数据样本设计、模型微调/调优到最终结果输出的全生命周期开发流程。
End-to-end ownership of the development lifecycle, from data sample design and model fine-tuning/optimization to final output delivery.

Requirements

任职要求 / Requirements
1.本科及以上学历，计算机科学、人工智能或相关专业背景；拥有 3 年以上大模型应用或算法开发经验。
Bachelor’s degree or higher in Computer Science, AI, or related fields; 3+ years of experience in LLM application development or algorithmic research.
2.熟练掌握 Python 或 TypeScript，具备独立编写复杂结构化数据解析脚本与 LLM 接口调用代码的能力。
Proficiency in Python or TypeScript, with the ability to independently write complex structured data parsing scripts and LLM integration code.
3.深入理解 HTML DOM 结构，熟练运用 CSS Selector、XPath 等文本定位与解析技术。
Deep understanding of HTML DOM structures and proficiency in text localization and parsing techniques such as CSS Selectors and XPath.
4.精通 OpenAI API、Claude API 以及 LangChain 或 LlamaIndex 等主流框架，具备多模型接入与切换的实战经验。
Expertise in OpenAI/Claude APIs and frameworks like LangChain or LlamaIndex, with hands-on experience in multi-model integration.
5.具备优秀的数据清洗、语义抽取与文本规范化经验，熟练使用 BeautifulSoup、正则表达式及 Prompt 流程优化工具。
Strong background in data cleaning, semantic extraction, and text normalization using BeautifulSoup, Regex, and prompt workflow optimization.
6.能够独立完成从样本准备、提示词工程调优到批量生产输出的完整业务流程。
Proven ability to independently manage the complete workflow from sample preparation and prompt tuning to batch production.
7.拥有稳定的网络连接环境，具备极强的代码规范意识和版本控制习惯。
Reliable high-speed internet connection and a strong commitment to clean code standards and version control best practices.
8.良好的英语读写能力，能够无障碍阅读前沿技术文档并利用国际开发者平台资源解决问题。
Proficient in English, capable of reading technical documentation and leveraging global developer resources to solve complex problems.

加分项 / Bonus Points
1.拥有成熟的 LLM 应用落地经验，特别是在自动化解析、智能问答或复杂数据抽取领域。
Proven track record of deploying LLM applications, especially in automated parsing, Q&A systems, or data extraction.
2.熟悉 Cloudflare Workers、Vercel 或 AWS Lambda 等 Serverless 无服务器架构。
Familiarity with Serverless architectures such as Cloudflare Workers, Vercel, or AWS Lambda.
3.对 AI Agent、RAG（检索增强生成）及知识提取体系有深入研究或浓厚兴趣。
In-depth research or strong interest in AI Agents, RAG (Retrieval-Augmented Generation), and knowledge extraction systems.
4.具备大模型微调（Fine-tuning）、推理加速优化或大规模分布式调用经验。
Experience in LLM fine-tuning, inference optimization, or large-scale distributed system design.
5.在 NeurIPS、ICML、ACL 等顶级学术会议发表过论文，或活跃于 Hugging Face 等开源社区者优先。
Publications in top-tier conferences (NeurIPS, ICML, ACL) or active contributions to open-source communities like Hugging Face are highly preferred.

Benefits

薪酬福利：
1.依据评估与交付范围协商；若交付质量出色，可长期合作并共同参与产品发布迭代。
2.项目制结算（按阶段/里程碑支付）+ 弹性工作时间 + 远程工作。
Compensation and Benefits
Negotiable based on project scope and evaluation; potential for long-term collaboration and participation in product release iterations for high-quality deliveries.
Project-based settlement (paid by phase/milestone) + flexible working hours + remote work.