生成式AI专员
Gen AI Specialist
职位描述：
Gen AI 专家是音频视觉生成式 AI 领域的专家，积极参与相关社区。他们负责识别和评估前沿技术与工作流程，涵盖开源和闭源生态系统，并独立原型化符合公司需求的 POC 实现。
该团队位于研究与产品/创意团队的交汇处。作为研究小组的应用延伸，他们参与研究冲刺，作为内部开发实现和测试的第一接触点。利用其领域专业知识以及来自产品和创意合作者的见解，为研究部门提供有价值的情境信息和有根据的新方向建议。我们正在 EMEA 地区远程招聘。
主要职责
- 在开源和闭源生态系统中识别和评估前沿 GenAI 技术和工作流程。
- 原型化符合公司需求的 POC 实现。
- 作为内部研究开发实现和测试的第一接触点。
- 通过利用产品和创意洞察，为研究团队提供有价值的情境信息和有根据的建议。
- 构建可扩展的 ComfyUI 工作流，并帮助维护一个共享、不断演进的 ComfyUI 环境。
- 作为跨学科团队的一员，推动项目和产品的交付。
- 清晰地向开发者和创意人员展示研究成果，突出优势、劣势和示例。
必备条件
- 有测试和训练逼真生成模型的专业经验，最好专注于视频应用。
- 能够搭建模型推理和训练的测试环境。必须熟悉 Conda、git、依赖管理，并具备足够的 Python/PyTorch 知识，能够根据具体需求调整代码。
- 有将现有工具和技术有效整合到更广泛的定制管道和工作流中的实际经验。
- 能够管理大型训练数据集。深入了解标注、筛选和增强数据对训练模型的影响。
- 能够调整模型调参，理解这些参数对训练模型的影响。
- 有与开发者协作的经验（研究团队合作经验优先）。
- 非常熟练使用 ComfyUI。能够构建可扩展的工作流，在需要时创建自定义节点，并维护环境。
查看英文原文
Job Description:
Gen AI Specialists are experts in Audiovisual Generative AI with active engagement in its surrounding communities. They are responsible for identifying and evaluating cutting edge technologies and workflows, across both open and closed-source ecosystems, and independently prototyping POC implementations aligned with the company’s needs.
This team operates at the intersection of research and product/creative teams. Functioning as an applied extension of the research group, they follow research sprints serving as the first point of contact for implementing and testing in-house developments. While leveraging their domain expertise, and insights from product and creative collaborators, to offer the research department valuable context and well-informed proposals for potential new directions. We are hiring remotely across the EMEA region.
Key Responsibilities
- Identify and evaluate cutting-edge GenAI technologies and workflows across open and closed-source.
- Prototype proof-of-concept implementations aligned with company needs.
- Act as the first point of contact for implementing and testing in-house research developments.
- Offer research teams valuable context and informed proposals by leveraging product and creative insights.
- Build scalable ComfyUI workflows and help maintain a shared, evolving ComfyUI environment.
- Work as part of a multi disciplinary team to deliver across projects and products.
- Clearly present research findings to developers and creatives, highlighting strengths, weaknesses, and examples.
Must Haves
- Professional experience testing and training photoreal generative models, ideally with a focus on video applications.
- Comfortable setting up testing environments for model inference and training. Must be confident with Conda, git, dependency management, and have enough Python/PyTorch knowledge to adapt code to specific requirements.
- Proven experience at effectively implementing existing tools and techniques into wider custom pipelines and workflows.
- Comfortable managing large training datasets. Deep understanding of the effects captioning, curating and enhancing them has on the trained model.
- Comfortable tweaking model tuning hyperparameters, with an understanding of the effects these have on the trained model.
- Experience collaborating with developers (research teams as a bonus).
- Very proficient in ComfyUI. Comfortable building scalable workflows, creating custom nodes where needed, and maintaining a shared, evolving ComfyUI environment.
- Deep engagement and awareness of the Gen AI communities and offerings, across both open and closed-source spaces.
- Comfortable independently researching specific topics, gathering and clearly presenting findings to both developers and creatives - highlighting key strengths, weaknesses, backed by relevant examples.
- Comfortable onboarding and training artists from other disciplines to expand the wider Gen AI artist group.
Nice to Have
- Prior role in a research or RnD-focused environment.
- Professional background either as a creative or a developer in a visual field such as VFX, visual arts, or a visually-driven tech company.
Originally posted on Himalayas