Jobs · Remotely
LLM Agent Research Scientist
Canva · Beijing · Posted today
Available in 2 locations
Beijing · hybrid Apply → Beijing, Beijing, China · hybrid Apply →About the role
Define and lead LLM agent research initiatives and develop agent architectures and frameworks.
📋 Description Define and lead LLM agent research initiatives Design agent architectures, reasoning, planning, tool-use, and multi-agent frameworks Develop post-training recipes and RL methods for agent capabilities Train models on large high-quality datasets with high-end GPUs Build agent harness for reliable, complex long-horizon tasks Create benchmarks and evaluations for agent correctness and safety 🎯 Requirements Hands-on experience developing LLM-based agents or foundation models Post-training, RL, reward modeling, synthetic data, agent environment design Publications at top venues or impactful open-source contributions Proficient in Python and PyTorch Knowledge of Transformers, DeepSpeed, Megatron, agent/RL frameworks Cloud computing for efficient training and deployment
Read the full posting on Remotely →
FAQ
Is the LLM Agent Research Scientist role at Canva remote?+
This LLM Agent Research Scientist position is listed as hybrid (Beijing).
What seniority level is this LLM Agent Research Scientist role?+
This is a mid level position.
How do I apply for the LLM Agent Research Scientist role at Canva?+
Use the "Apply on Remotely" button to open the original posting on Remotely, where you can submit your application directly to Canva.