Advance LLM learning through reinforcement learning research at Huawei Canada’s Noah’s Ark lab. This 12-month contract focuses on interactive environments. The Reinforcement Learning Researcher will develop research for continual, agentic self-improvement beyond static fine-tuning. Work spans deep learning, Transformers, language model post-training, and interactive agents. You will implement and evaluate training and benchmarking pipelines for reasoning-enhanced LLMs and tool-using systems. Key Responsibilities:
Enable LLMs to learn from experience and workplace feedback Study RLHF, GRPO, and reward-free post-training methods Develop reinforcement learning for tool-using and browsing agents Design multi-turn,
verifiable reasoning evaluation tasks Scale experiments across large GPU clusters and publish insights Requirements:
PhD in Computer Science or related field, or comparable master’s experience Strong deep learning foundation, including Transformers and optimization Research experience in reinforcement learning or language model fine-tuning First-author paper at a listed top-tier AI venue Python, PyTorch, DeepSpeed, and distributed training proficiency Apply reinforcement learning, distributed training, and evaluation expertise to advancing reasoning-enhanced LLM research at Huawei Canada.