LLM Agent Research Scientist

About the Role

As an LLM Agent Research Scientist youll define and lead research that pushes the limits of what LLM-powered agents can do. Youll design agent architectures reasoning and planning methods tool-use and multi-agent frameworks and the post-training recipes (including RL) that make agents capable and dependable. Youll also train models with high-end GPUs on large high-quality datasets. Youll sit at the intersection of frontier research and product directly shaping how LLM agents transform the way people create and get work done in Canva.

作为大模型智能体研究科学家你将定义并主导不断突破智能体能力上限的研究你将设计智能体架构推理与规划方法工具调用与多智能体框架以及让智能体更强大更可靠的后训练方案包括 RL你也会在大规模高质量的数据集上用高端 GPU 训练模型你将处在前沿研究与产品的交汇点直接影响大模型智能体如何改变人们在 Canva 上创作与工作的方式

At the moment this role is focused on

• Research Direction: Determining and leading agent research initiatives aligned with the frontier of LLM/agent research and Canvas product goals.

• Model Innovation: Creating novel agent capabilities through post-training and RLdeveloping reward modeling synthetic data and environment designand running experiments with high-end GPUs on large high-quality datasets to validate hypotheses and translate results into shippable improvements.

• Harness Innovation: Designing and building the agent harness that lets agents execute complex long-horizon tasks reliably in production and exploring harness techniques for self-improvement.

• Evaluation: Creating rigorous benchmarks and evaluation methods for agent correctness safety robustness and real-world task success.

• Product Impact: Partnering closely with engineering product design and data teams to land frontier research capabilities in real Canva product experiences.

• 研究方向定义并主导契合大模型/智能体研究前沿与 Canva 产品目标的研究方向

• 模型创新通过后训练和 RL 打造全新的智能体能力涵盖 reward modeling合成数据环境设计等并在大规模高质量数据集上用高端 GPU 开展实验验证假设把结果转化为可落地的改进

• Harness 创新设计并构建 agent harness让智能体能在生产环境中稳定地完成复杂的长链路任务并探索让智能体自我改进的 harness 技术

• 评测围绕智能体的正确性安全性鲁棒性和真实任务完成情况构建严谨的基准与评测方法

• 产品影响与工程产品设计和数据团队紧密协作将前沿研究能力落地到真实的 Canva 产品体验中

Qualifications :

Youre probabaly a match if you have:

• Rich hands-on experience developing and iterating on LLM-based agents or foundation modelscovering areas such as post-training RL reward modeling synthetic data or agent environment design.

• A strong academic and professional track record including peer-reviewed publications at top-tier venues such as NeurIPS/ICML/ICLR/ACL/EMNLP/CVPR and/or impactful open-source contributions.

• Proficiency in Python and PyTorch with familiarity in frameworks and libraries such as Transformers DeepSpeed Megatron and agent/RL frameworks as well as cloud computing platforms for efficient training and deployment.

• Curiosity! Always looking to stay ahead of industry trends by tracking AI literature competitors and emerging technologies.

• 在研发和迭代大模型智能体或基础模型方面有丰富的实战经验涵盖后训练RLreward modeling合成数据或智能体环境设计等方向

• 扎实的学术与专业履历例如在 NeurIPS/ICML/ICLR/ACL/EMNLP/CVPR 等顶级会议发表过学术论文或有过有影响力的开源贡献

• 精通 Python 和 PyTorch熟悉 TransformersDeepSpeedMegatron 及 agent/RL 框架等工具库以及用于高效训练和部署的云计算平台

• 好奇心始终关注行业动态通过追踪 AI 文献竞品和新兴技术保持领先

Additional Information :

该岗位现面向所有经验阶段的候选人开放包括社会招聘应届毕业生同时开放实习生岗位工作地点为北京欢迎申请期待你的加入

Notice: This position is open to candidates at all experience levels including experienced candidates graduates as well as internship opportunities. The role is based in Beijing. We welcome your application and look forward to having you on board!

Remote Work :

No

Employment Type :

Full-time