Chief Scientist - Foundation Models

About the role

Huawei Canada is seeking a Chief Scientist to lead the scientific strategy for foundation models, including large language and multimodal models. The role will shape a multi‑year roadmap, drive breakthrough research, and translate innovations into production‑grade capabilities.

Key responsibilities

• Define the scientific strategy and long‑term vision for foundation models, creating a multi‑year roadmap with milestones, investment priorities, and success metrics.
• Lead research excellence at scale by recruiting, mentoring, and retaining top talent while establishing rigorous standards for experimentation, evaluation, reproducibility, and model governance.
• Drive innovation in model architecture, large‑scale training, data strategy, optimization, inference, evaluation, and alignment, balancing frontier exploration with reliable deployment.
• Own the end‑to‑end research portfolio, prioritize bets across exploratory and applied research, review progress, unblock teams, and ensure high‑impact delivery.
• Translate research into production‑grade foundation model capabilities by partnering with engineering and product leaders to meet performance, cost, latency, safety, and reliability goals.

Required profile

• PhD in Machine Learning, AI, Computer Science, or equivalent demonstrated scientific leadership.
• Internationally recognized track record in foundation models with sustained contributions to top‑tier venues or widely adopted open‑source work.
• Deep technical mastery of large‑scale training, data curation, evaluation, alignment, safety, inference optimization, and deployment constraints.
• Proven senior leadership experience managing principal researchers and multiple teams, setting direction and raising standards.
• Demonstrated ability to turn novel research into robust, measurable business and user impact.
• Strong communication and executive presence to represent the company externally and influence internal stakeholders.

Required skills

• Large‑scale model training
• Data curation for foundation models
• Evaluation methodologies for LLMs
• Model alignment and safety
• Inference optimization
• Deployment constraints