Machine Learning Researcher – LLM Agents & Efficient Deep Learning

About the role

Huawei Canada is seeking a Machine Learning Researcher for a 12‑month fixed‑term position in Montreal. The role focuses on advancing large language model (LLM) agents and efficient deep‑learning techniques within the Noah’s Ark Laboratory.

Key responsibilities

• Design and develop multi‑agent systems and function‑call models for LLM agents.
• Build scalable SLM pipelines for training, evaluation, and deployment.
• Research and implement efficient training and inference methods such as quantization (INT8/FP8), pruning, sparsity, and low‑rank tensor decompositions.
• Explore parameter‑efficient fine‑tuning techniques (e.g., LoRA) and long‑context memory modeling.
• Adapt GPU‑based LLMs to CPU‑based SLMs and optimize inference performance (latency, throughput, KV‑cache efficiency).
• Prototype new ideas, validate through rigorous experimentation, and integrate reasoning models with tool‑use frameworks (function calling, APIs).

Required profile

• MSc or PhD in Computer Science, Electrical Engineering, or a related field.
• Strong publication record at top AI conferences (NeurIPS, ICML, AAAI, ICLR, etc.).
• Deep‑learning and optimization expertise, with experience in transformer architectures, LLMs and SLMs.
• Proficiency in PyTorch and solid understanding of linear algebra, probability, and optimization algorithms (SGD, Adam).
• Demonstrated ability to implement and scale ML systems in research or production environments.

Required skills

• PyTorch
• Transformer architectures
• Large Language Models (LLM) and Small Language Models (SLM)
• Quantization (INT8/FP8)
• Pruning and sparsity
• Low‑rank and tensor decomposition
• LoRA fine‑tuning
• GPU and CPU deployment
• Optimization algorithms (SGD, Adam)

What we offer

• Target annual compensation between $106,000 and $156,000 (CAD), based on education and experience.
• Inclusive and accessible recruitment process with accommodations available.