Machine Learning & Reinforcement Learning Engineer

About the role

Huawei Canada's Software-Hardware System Optimization Lab is seeking an Engineer to design and build scalable infrastructure for reinforcement learning, online search, recommendation systems, and large model fine‑tuning over a 12‑month contract.

Key responsibilities

• Design and build scalable infrastructure to support RL, online search, recommendation systems, and large model fine‑tuning and deployment.
• Develop efficient ML solutions for recommendation systems and RL problems, including multi‑armed and contextual bandits, tree search, and multi‑agent orchestration.
• Implement and optimise deep learning architectures, such as custom Transformers for agentic decision‑making systems.
• Apply search and optimisation techniques to efficiently fine‑tune RL and ML models.
• Work with large multimodal models (LLMs, VLMs), analyse components, and fine‑tune for task‑specific applications.
• Conduct systematic benchmarking, literature review, experimentation, and validation in simulation and real‑world environments.
• Collaborate with research teams to scale online RL training capabilities and improve system robustness and accuracy.
• Explore and integrate emerging AI methodologies and tools into production platforms.

Required profile

• Master’s or PhD in Computer Science, Machine Learning, or a related field.
• Excellent Python programming skills with strong software engineering practices.
• Strong foundation in Reinforcement Learning, Deep Learning, Recommender Systems, and Transformer‑based architectures.
• Demonstrated experience implementing RL algorithms beyond academic prototypes.
• Hands‑on experience with PyTorch and distributed training frameworks such as DeepSpeed.
• Proven research excellence, including at least one publication in top‑tier venues (e.g., NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, ICRA, RLC).

Required skills

• Python programming
• PyTorch
• DeepSpeed (distributed training framework)
• Reinforcement Learning algorithms
• Transformer architectures
• Large Language Models (LLMs) and Vision‑Language Models (VLMs)