Skip to content

PUBLIC ROLE INFORMATION

资深 Agent 工程师(Tripo Studio · Canvas Agent)

Tripo北京

Key conditions

Locations
北京
Employment
未确认
Track
Experienced

Not disclosed: Team, Salary, Employer application deadline

Sources and available information

Reported open · Last observed 2026-10-18

Topics
Unknown
Company
Tripo · Source supported · Coverage unknown
Work mode
Unknown
Source type
官方岗位页 · Source supported · Coverage unknown
Company type
创业公司 · Tentative · Coverage unknown
Hiring track
Unknown
Remote scope
Unknown
Role
开发 · Source supported
Business / team
Unknown
Location
北京 · Source supported
Employment
Unknown
Role AI context
直接研发/交付AI · Source supported
Company AI context
Unknown
Recruitment status
来源显示在招 · Source supported

Unknown information is retained as such and updated when new source information is available.

Application path and records

Apply through the official channel. The full external receipt stays in your workspace. When signed in, three minimal event types explicitly recorded through the CLI are reported by default: selection, self-reported application and channel observation. Browsing or saving a role does not report an event; external applications do not appear under My applications.

Role description from source

【我们在做什么】
Tripo 做 3D 生成:把文本、图片和交互变成可交付的 3D 资产。https://tripo3d.ai

本岗位在 Agent 团队,负责 Studio 中的 Canvas Agent:把自然语言、参考图和画布操作,编排成可执行的创作流程。

工作重点是 Agent 架构、评测,以及效果、成本、时延之间的取舍。

【你将负责】
1. 端到端创作链路:基于自然语言、参考图和画布交互,规划并执行「生成 → 检查 → 分割 / 局部编辑 → 贴图 → 重拓扑 → 绑骨 → 导出」,交付可恢复、可评测的生产级资产。
2. Agent 架构演进与工具化:把 Studio 已有的生成与后处理能力,沉淀为稳定、可组合的 Tool / Skill / MCP;演进编排与控制循环,定义接口契约、权限边界、计量计费和失败语义。
3. 画布上下文:设计画布侧状态协议,把场景图、部件选中、相机位姿、编辑历史与当前任务纳入 Agent 决策,并与对话上下文统一编排。
4. 长程任务状态机:覆盖提交、排队、预览、取消、重试、换种子重跑和断点恢复;处理超时、部分成功和失败恢复,而不是只对接同步 API。
5. 效果、成本与时延的取舍:这是一组持续 trade-off,不是单向追精度。按资产用途在快速网格与高精度模型之间路由,抑制无效重试;依据模型能力边界做选型、工具划分和停止条件,而不是默认调用最强模型。
6. 评测、triage 与效能:建设离线评测集和在线指标;对失败与低采纳 case 做 triage,把调优做成可重复迭代,持续提升效果、成本和时延上的效能。用户采纳、重生成、局部修改和导出用途,作为数据飞轮输入。
7. 跨团队落地:与前端、后端、算法、评测和 Infra 对齐接口、状态协议与交付节奏,把 Canvas Agent 稳定推到线上。

职位要求
【任职要求】
1. 本科及以上学历,计算机相关专业;3 年及以上后端或平台开发经验,能独立定位线上故障并做性能治理。
2. 精通 Go,有服务端系统设计与落地经验;能独立完成高并发与异步任务架构。熟悉缓存、消息队列、限流、熔断、降级与灰度;熟悉 Docker / Kubernetes、PostgreSQL、Redis。
3. 有 AI / Agent 相关的生产实践:做过线上质量、工作流、模型调用或同类系统即可。掌握意图理解与澄清、工具调用(Function Calling / MCP / Skill)、记忆与上下文压缩;能独立讲清并实现 Agent Loop,不绑定单一框架。
4. 理解 Agent 控制循环与系统拓扑的分工。熟悉 ReAct、Plan-Execute、Critic / Reflection 等循环,能判断观察是否改写下一步、何时打回、何时终止,并区分探索循环与步骤已锁定的定稿生成。熟悉 Router / Pipeline / Orchestrator 等拓扑,能设计产物与权限在节点间的交接,避免循环正确但交接失败、整链重跑。
5. 有 Agent Infra 生产实践:能落地模型网关、多模型路由、用量与成本控制、链路追踪与回放。能讲清 Runtime 在生产里承担什么、不承担什么,不把「会调框架 API」当成 Runtime 能力本身。
6. 能对线上坏结果做 triage,定位是规划、工具、模型、上下文还是停止条件的问题;能据此调优,并用 eval 和效能指标验证,而不是只看链路是否跑完。持续跟踪前沿模型与 Agent 工程实践。

【加分项】(按匹配度)
1. 熟悉主流 Agent Runtime 框架(如 Eino、Google ADK),并有生产使用经验。
2. 有 2D / 3D 创作 Agent,或图像 / 视频 / 3D 生成工作流编排经验(如 ComfyUI、DCC 插件、节点式画布)。
3. 有 VLM / 多模态 Agent 经验,能基于视觉反馈驱动下一步编辑。
4. 有 MCP / Skill / CLI 设计经验,能向其他 Agent 或 IDE 暴露稳定工具协议。
5. 有 Agent 评测集建设,或把用户行为闭环为数据飞轮的实践。
6. 有向量检索、会话记忆或项目级记忆的工程实践。

ABOUT THE COMPANY / TEAM

Tripo

Startup

Tripo Studio、3D生成API、创作Agent与3D资产生态

Company website ↗

Other open jobs at this company

Tripo北京, 杭州Engineering

Full Stack Engineer全栈工程师(Tripo Studio)

Reported open · Last observed 2026-10-18

Your Responsibilities 1.You will be granted full ownership and end-to-end accountability for the design and delivery of core product modules: Full-Stack Product Development Backend: Design and build high-concurrency, low-latency core services with Go/Python/Node.js, covering task scheduling, billing systems, authentication frameworks and asynchronous workflows. Frontend: Independently develop user-facing web applications including creation workspaces, data dashboards, online 3D preview and interactive playgrounds, with a relentless focus on premium user experience. Full lifecycle ownership: You take full control over the entire tech stack, from database schema design and API contracts to CI/CD pipelines and production monitoring. 2.Integration & Innovation of AI Capabilities Deeply embed AI model capabilities (3D generation, post-processing, semantic comprehension) into product workflows, and design optimal interaction paradigms between users and AI. Leverage LLMs and Agent tools to boost engineering efficiency: automated test generation, code review, documentation generation and intelligent operations. Keep track of cutting-edge AI technologies (multimodal models, Diffusion, NeRF/

Tripo杭州Engineering

Senior Agent Development Engineer 高级 Agent 开发工程师

Reported open · Last observed 2026-10-18

Design, develop and launch AI Agents for real‑world business workflows. Deliver production‑grade systems that serve real end‑users and continuously undertake business tasks, and take ownership of business outcomes. Iterate continuously based on online task performance and user feedback. Improve task success rate, stability, response latency and cost through root‑cause analysis of failure cases, evaluation regression testing and execution trace troubleshooting. Build core Agent capabilities including task planning, tool calling, context management and memory module. Integrate business systems and data, and resolve challenges such as exception recovery, state management and permission control for complex tasks. Select and extend Agent frameworks according to business requirements, and accumulate reusable engineering capabilities and tuning best practices. 1、面向真实业务流程,负责 AI Agent 的设计、开发与上线,交付有实际用户使用、能够持续承担业务任务的系统,并对业务效果负责; 2、围绕线上任务和用户反馈持续迭代,通过失败案例分析、评测回归及执行轨迹排查,优化任务成功率、稳定性、响应速度与成本; 3、负责任务规划、工具调用、上下文管理、Memory 等核心能力建设,打通业务系统与数据,解决复杂任务中的异常恢复、状态管理和权限控制问题; 4、结合业务需求选择和扩展 Agent 框架,沉淀可复用的工程能力与调优方法。 职位要求 Hands‑on experience building production‑grade business‑facing Agents that have been launch

Tripo北京, 上海Engineering

Senior SRE Engineer (AI Infrastructure)高级 SRE 工程师 (AI-INF-基础设施)

Reported open · Last observed 2026-10-18

Job Responsibilities 1.Multi-cloud Architecture Management & Business Implementation Take charge of the planning, construction and daily operation & maintenance of the company’s infrastructure on mainstream public clouds including AWS and Alibaba Cloud. Independently collaborate with business teams to deliver architecture design, resource planning, deployment & launch, and full-lifecycle management for complex business systems. 2.Stability Assurance of Kubernetes Clusters Build large-scale Kubernetes clusters, optimize their stability, conduct capacity planning and fine-tune scheduling policies. Lead containerization migration of services, and troubleshoot complex issues of cloud-native components such as network and storage. 3.O&M of AI Computing Infrastructure Ensure stable operation of large model training and inference workloads. Be proficient in driver configuration, network (InfiniBand/RoCE) and monitoring troubleshooting of heterogeneous computing servers equipped with NVIDIA GPUs, and optimize GPU resource scheduling and utilization efficiency. 4.CI/CD & Automated Pipeline Construction Design and optimize CI/CD pipelines based on GitHub Actions, GitLab CI, ArgoCD, etc. Prom

Tripo北京Product

Search&Recommendation Strategy Product Manager 搜索推荐策略产品经理(UGC社区方向)

Reported open · Last observed 2026-10-18

1. Design search and recommendation strategies for Tripo Studio, building a content discovery and distribution system for the UGC creator ecosystem and improving content relevance and user experience. 2. Optimize retrieval, ranking, recommendation, traffic allocation, and cold-start strategies for 3D and UGC content to improve content quality, exposure, and creator growth. 3. Develop strategies for search ranking, personalized recommendations, content quality, and distribution efficiency to improve user engagement and retention. 4. Build data analysis frameworks, track key metrics, identify optimization opportunities, and drive continuous improvements through experiments. 5. Collaborate with algorithm, engineering, design, operations, and business teams to launch and iterate search and recommendation solutions. 6. Follow trends in generative AI, 3D content, and creator ecosystems to explore new content discovery and distribution approaches. 1. 负责 Tripo Studio 内容搜索与推荐策略产品设计,搭建面向 UGC 创作者生态的内容发现与分发体系,精准匹配用户创作需求与优质3D内容资产,提升内容分发效率和用户体验。 2. 基于创作者生态与内容消费场景,优化3D内容及UGC内容的召回、排序、加权分发、冷启动及流量扶持策略,提升优质内容曝光效率,促进创作者成长和平台内容供给增长。 3. 负责搜索排序、推荐分发、内容召回、流量分配、去重降噪等核心策略机制设计,解决内容同质化、低质内容泛滥、流量分配不均等问题,持续优化用

Tripo杭州Product

User Experience Product Manager 用户体验产品经理

Reported open · Last observed 2026-10-18

岗位概述 Responsible for building the next-generation AI generative POD e-commerce platform. Leveraging AI to lower creative barriers, we cater to the demand for customized products from consumers, lifestyle enthusiasts, designers, and 3D printing hobbyists. You will balance community ecosystem development with the entire lifecycle management of e-commerce, creating a closed-loop experience from "inspiration generation" to "physical delivery.". 负责构建下一代AI生成式POD电商平台。通过AI降低创作门槛,满足消费者,生活家、设计师及3D打印爱好者对定制化产品的需求。您需平衡社区生态建设与电商全生命周期管理,打造从“灵感生成”到“实体交付”的闭环体验。 岗位核心职责 1. Responsible for the establishment of user activation, growth tiering, and retention systems. Through data mining and analysis, identify pain points in user activation, engagement, and retention, and continuously optimize key conversion paths. 2. Deeply understand personalized consumption and creative pathways, integrate user research with data analysis to accurately identify core issues and drive product iteration, ensuring a closed-loop experience. 3. Possess foundational product design capabilities, assist users in understanding the value of AI-generated models and physical POD services, and guide them in developing long-term cr