Intern Research Engineer - AI Agent Systems

Huawei Technologies Canada Co., Ltd.•Markham, ON

1d•CA$58,000 - CA$104,000

About The Position

Huawei Canada has an immediate internship opening for a Research Engineer. The Technology Planning and Cooperation Department promotes strategic innovation across all of Huawei Canada. Working with fellow experts, the focus is researching new and upcoming areas of technology as well as strategic planning to help achieve the long-term vision of Huawei. This department offers a unique opportunity to leverage both technical and business skills.

Requirements

Currently pursuing a Master’s or Bachelor’s Degree in Computer Science or a highly quantitative field, with a focus on AI, Software Engineering Systems, or Distributed Systems.
Proficient in orchestrating LLMs programmatically.
Deep understanding of Agentic design patterns (Planning, Memory, Tool Use, Multi-Agent Collaboration).
Exceptional coding skills in Python and/or TypeScript/Node.js.
Familiarity with structural design patterns and engineering clean, maintainable codebases.

Nice To Haves

Contributions to open-source AI Agent frameworks, developer tools, or compiler-related projects is an asset.
Experience with Code-LLM optimization (e.g., custom fine-tuning for code generation, structured outputs), embedded database engines or high-performance vector stores is an asset.
Familiarity with cross-framework mobile compilation or bridge layers (e.g., Android/AOSP or hybrid application runtimes) is an asset.

Responsibilities

Architect and implement high-scalability AI Agent pipelines, transforming complex requirements into reliable execution structures (e.g., TaskGraph IR, Operator DAGs, and multi-agent coordination frameworks).
Develop and optimize automated code generation, reasoning, and self-debugging loops that interact directly with multi-framework environments (e.g., creating bridge layers, compilation pipelines, or automated code refactoring tools).
Design and optimize advanced memory substrates for Agents, tackling challenges related to long-context handling, state persistence, and execution efficiency.
Establish rigorous, automated evaluation frameworks to benchmark Agent execution accuracy, latency, and reliability under non-deterministic LLM behaviors.
Translate the latest research in Agentic workflows (e.g., ReAct, Reflection, Planning, Tool Use) into production-grade, highly performant software libraries.