Multimodal AI Research Intern: Audio & Agents

Huawei Canada; Markham; On-site; CAD 58,000 - 104,000; Full time

Job summary

Huawei Canada is seeking an Audio, Speech, & Acoustics Research Intern to join our AI & Signal Processing Research team. You’ll explore, design, and implement cutting-edge algorithms at the intersection of acoustic signal processing, deep learning, and speech processing.

The ideal candidate is pursuing a Master’s or Ph.D. in CS, with strong Python and PyTorch experience, and a track record of research activity and publications.

Qualifications

• Pursuing a Master’s or Ph.D. in CS or related field.
• Strong foundation in AI, ML, and signal processing.
• Experience with Python and PyTorch in research settings.
• Publications in top venues (NeurIPS, ICML, CVPR, etc.) is a plus.

Responsibilities

• Conduct research on AI agents, LLMs, VLMs, code generation, and multimodal AI.
• Design and develop GUI agents and coding agents for web, mobile, and computer-use environments.
• Develop SFT, reinforcement learning, agent training, tool use, and function calling approaches.
• Build research prototypes, datasets, training pipelines, benchmarks, and evaluation frameworks.
• Publish findings in top-tier venues and communicate results through talks/publications.

Skills

Python

PyTorch

LLMs

VLMs

multimodal learning

reinforcement learning

generative AI

code generation

AI agents

computer vision

NLP

GUI agents

web agents

mobile agents

publications

Education

Master's or Ph.D. in Computer Science or related

Tools

Playwright

HTML

CSS

JavaScript

browser APIs

Job description

Huawei Canada is seeking an Audio, Speech, & Acoustics Research Intern to join our AI & Signal Processing Research team. You’ll explore, design, and implement cutting-edge algorithms at the intersection of acoustic signal processing, deep learning, and speech processing.

The ideal candidate is pursuing a Master’s or Ph.D. in CS, with strong Python and PyTorch experience, and a track record of research activity and publications.