职位介绍：负责规划和构建公司的核心数据资产体系，带领用户画像团队深入挖掘用户行为数据，并将数据转化为可复用的资产，支撑 AIGC、影像智能体及推荐算法的快速迭代。主导在推荐、订阅等高价值场景的验证，确保数据资产在业务增长和模型进化中发挥决定性作用。

岗位职责:

1、数据资产体系构建：负责「用户数据资产」与「训练数据资产」的沉淀，构建适配大模型微调（Fine-tuning）和智能体长短期记忆（Memory）的高质量数据集。

2、画像团队管理与资产规划：带领用户画像团队，统筹全公司级标签体系（Tag System）的设计与维护，确保画像资产的准确性、实时性及跨业务线的复用价值。

3、高价值场景落地验证：主导数据资产在推荐系统、订阅转化、Agent 交互等场景的应用，通过数据闭环验证模型效果，驱动业务增长。

4、全链路数据治理与 AI 赋能：利用 AI 技术提升画像建模与数据监控的效率，实现从“原始日志”到“结构化画像”再到“业务 Action”的自动化流转。

5、跨团队协同与项目推动：作为核心推手，协同算法、产品及运营团队，将业务挑战转化为数据方案。展现极强的主动性，确保跨部门项目高效达成。

任职要求：

1、经验背景：五年以上互联网数据架构或用户画像相关经验，有影像类、工具类 App 或 AI 场景背景。具备一定的团队管理经验，能激发团队创造力。

2、画像建模深度：深刻理解用户行为建模，熟悉用户分层、流失预警、生命周期管理及审美/偏好特征量化等实战方法。

3、AI 技术洞察：深刻理解大模型基础原理，了解如何将用户画像作为 Prompt 约束或 RAG 检索条件来优化 AI 生成效果；关注 AIGC 最新技术进展。

4、硬核技能：精通 SQL，熟悉 Spark 进行大规模数据清洗、统计分析及建模预研；熟练掌握 Python。

5、AI 工具力：熟练使用各类 AI 编程辅助工具（如 Cursor、Claude Code 等），能利用 AI 提升团队研发效能。

6、主动性与沟通：具备极强的Owner意识和跨团队推动能力，能向非技术方清晰解释数据资产的业务价值。

7、方法论：熟悉数据分析方法论、A/B 实验理论及数据血缘管理。

The Challenge Ahead:

At Meitu, data isn’t just a byproduct — it’s the raw material that drives our AIGC models, imaging agents, and recommendation engines. We’re looking for a Data Asset Architect to own the end-to-end data asset system that turns raw user signals into structured, reusable intelligence. You’ll lead a user profiling team, build datasets fine-tuned for LLMs and agent memory, and validate your work in high-stakes loops like recommendation and subscription conversion. This is a role for someone who thinks in data flows, loves clean architecture, and wants to see their work directly shape AI-powered products used by millions.

How You’ll Contribute:

· Architect the company’s core data asset system, framing both user data assets and training data assets that produce high-quality datasets for large-model fine-tuning and agent short/long-term memory;

· Lead the user profiling team to design and govern a cross-business tagging system, keeping profiles accurate, real-time, and genuinely reusable — not just theoretically, but in production;

· Drive validation in high-value scenarios like recommendation, subscription conversion, and agent interaction, using closed-loop data to evaluate model effectiveness and guide business growth;

· Apply AI techniques to level up profiling, modeling, and data monitoring, and build automated pipelines that move from raw logs to structured profiles to business actions;

· Act as the connective tissue across algorithm, product, and ops teams, turning ambiguous business questions into concrete data strategies and pushing projects forward with a strong sense of ownership.

What We Hope You Have:

· 5+ years of hands-on experience in data architecture or user profiling, ideally in imaging apps, utility tools, or AI-driven environments; you’ve led or mentored a small team and know how to spark creative thinking;

· Deep domain knowledge in user behavior modeling — you’ve built segmentation, churn prediction, lifecycle management, and even aesthetic/preference quantification in practice, not just in theory;

· Clear mental model of how LLMs and AIGC systems work under the hood; you understand how to inject user profiles as prompt constraints or RAG retrieval filters to improve generation quality, and you track the fast-moving AIGC research landscape;

· Strong technical spine: you write SQL that can handle complex logic, use Spark for large-scale cleaning, stats, and prototyping, and are comfortable with Python for data work;

· You regularly use AI coding assistants (Cursor, Claude Code, or similar) to accelerate your own work and you push for their adoption to boost team productivity;

· Genuine owner mindset and the communication chops to explain data asset value to non-technical stakeholders — you bridge the gap between messy data and clear business intent;

· Solid grounding in data analysis methodology, A/B experiment design, and data lineage — you understand that reproducible insights and traceable data are what make a system trustworthy.