Senior Fullstack Architect (AI) - GP, Remote: Colombia - Costa Rica, Fulltime

免费注册后解锁"你的匹配分 vs 高分画像"与个人技能缺口。

- This position is open to candidates located in Colombia or Costa Rica only - We are looking for a Solutions Architect to lead the technical design of AdVise, a campaign-management platform used to operate approximately 1,600 Google Ads and Meta accounts for automotive dealerships. The role owns the architecture every capability depends on: the governed write path, the desired-state model, the precedence rules between systems, the permission model and the platform's operating envelope. The architect acts as the project's single design authority, works hands-on with the engineering team through the build, and leaves the client able to run and extend the platform after handover. This architect will also own the architecture and governance of the project's AI Harness, working with the Senior Fullstack Engineer who leads its implementation, and will make sure Release One anticipates the prescriptive and agentic capabilities planned for Release Two. Responsibilities Act as the single design authority for the platform, recording each decision in architecture decision records and the shared decision log. Define the provider-neutral desired-state and delta model and the governed write-path contract: validation, dry runs, execution, receipts, partial-failure handling, non-bypassable locks, and a live-state read before every write. Design the precedence model that decides which source wins when Salesforce, live platform state, playbooks, locks, compliance rules, inventory and offers disagree, together with the managed-scope and field-ownership model. Design the rules engine: rule classes, evaluation order, failure handling, validation on import, the forward gate at render and the backward sweep across live campaigns. Own the operating envelope: Google Ads and Meta rate limits, change-history polling cadence, a yield order between workloads, and headroom validated at full-book scale. Define the security and identity model, including roles, single sign-on, named service identities for every automated actor, privileged-action audit, and the approval authority matrix. Design the integration architecture and data contracts for Google Ads, Meta, Salesforce, Snowflake and the client's offer and incentive data, working with the client's data team. Choose the operational store, queueing and orchestration approach, and the build-versus-buy split for platform components on AWS, with the client's long-term maintainability in mind. Turn scale, latency, recovery and alert-volume expectations into measurable non-functional targets, and validate them through load and performance testing. Design the migration architecture with the Engineering Manager: adopt-in-place mechanics, rollback, the pilot wave scorecard, wave planning and acceptance tolerances. Own the architecture and governance of the AI Harness used to deliver the platform: guardrails, human review gates, approved tooling, spend controls and the client's security review. Design the Release Two runway, so that future AI agents carry their own identities and propose changes through the same governed write path, locks and audit trail as people do. Lead technical spikes, design reviews and code reviews, and review the architecture of every capability before its build starts. Take part in steering, write the technical content of milestone sign-off packages, document findings with proposed dispositions, and escalate technical blockers within one business day. Pair with the client's engineers throughout the build and prepare the documentation and decision records they need to operate and extend the platform independently. Collaborate with product leaders, paid-media strategists, the client's engineering and data teams, the Engineering Manager and the delivery team to translate complex requirements into a maintainable design. Technical Requirements Senior-level experience as a solutions or software architect on production platforms that integrate with external systems at scale. Experience designing

- 本职位仅面向位于哥伦比亚或哥斯达黎加的候选人 - 我们正在寻找一名解决方案架构师，负责领导 AdVise 的技术设计。AdVise 是一个广告活动管理平台，用于为汽车经销商运营约 1，600 个 Google Ads 和 Meta 账户。该职位负责所有能力所依赖的架构：受治理的写入路径、期望状态模型、系统之间的优先级规则、权限模型以及平台的运行边界。架构师作为项目的唯一设计权威，在构建过程中与工程团队 hands-on 协作，并在移交后让客户能够独立运行和扩展平台。该架构师还将负责项目 AI Harness 的架构与治理，与领导其实现的高级全栈工程师合作，并确保 Release One 预见 Release Two 中规划的规范性和代理性能力。

职责

作为平台的唯一设计权威，将每项决策记录在架构决策记录和共享决策日志中。

定义提供商中立的期望状态与增量模型，以及受治理的写入路径契约：验证、试运行、执行、回执、部分失败处理、不可绕过的锁，以及每次写入前的实时状态读取。

设计优先级模型，在 Salesforce、实时平台状态、playbook、锁、合规规则、库存和优惠之间出现分歧时决定哪个来源优先，同时设计受管范围与字段所有权模型。

设计规则引擎：规则类别、评估顺序、失败处理、导入时验证、渲染时的前向门控以及跨实时广告活动的后向扫描。

负责运行边界：Google Ads 和 Meta 的速率限制、变更历史轮询节奏、工作负载之间的产出顺序，以及在全量规模下验证的余量。

定义安全与身份模型，包括角色、单点登录、每个自动化参与者的具名服务身份、特权操作审计以及审批权限矩阵。

设计 Google Ads、Meta、Salesforce、Snowflake 以及客户优惠和激励数据的集成架构与数据契约，与客户的数据团队合作。

选择运营存储、队列与编排方案，以及 AWS 上平台组件的构建与购买划分，兼顾客户的长期可维护性。

将规模、延迟、恢复和告警量预期转化为可衡量的非功能性目标，并通过负载和性能测试进行验证。

与工程经理共同设计迁移架构：就地采用机制、回滚、试点波次记分卡、波次规划和验收容差。

负责用于交付平台的 AI Harness 的架构与治理：护栏、人工审核门控、经批准的工具、支出控制以及客户的安全审查。

设计 Release Two 的跑道，使未来的 AI 代理拥有自己的身份，并通过与人员相同的受治理写入路径、锁和审计轨迹提出变更。

领导技术 spike、设计评审和代码评审，并在每个能力的构建开始前审查其架构。

参与指导委员会，撰写里程碑签核包的技术内容，记录发现并提出处置建议，在一个工作日内上报技术阻塞。

在构建过程中与客户的工程师配对，并准备他们独立运营和扩展平台所需的文档和决策记录。

与产品负责人、付费媒体策略师、客户的工程和数据团队、工程经理以及交付团队协作，将复杂需求转化为可维护的设计。

技术要求

具备高级水平的解决方案或软件架构师经验，曾在生产平台上与外部系统大规模集成。

具备设计经验

⚑ 岗位信息收录自公开招聘页（Himalayas），以雇主官网为准