Machine Learning Engineer
Explicitly forbids vibe coding; focuses on low-level PyTorch model training and manual model engineering.
About the Role
Machine Learning Engineer to develop AI-powered brand identity and image-generation systems at a small Vancouver-based team, focusing on training LLMs and diffusion models, curating training data, and producing/validating graphic assets while supporting customers. The role is hybrid/onsite and requires hands-on PyTorch-level model work and vector graphic design skills.
Job Description
Role
We are hiring a Machine Learning Engineer to develop the next generation of an AI-assisted brand identity design chatbot and image-generation products. The role sits on a very small Vancouver-based team and requires wearing multiple hats across model training, data curation, design, and customer-facing tasks.
Key Responsibilities
- Train and fine-tune LLMs and diffusion/image models (LoRA and full-weight training).
- Scrape, hand-curate, and prepare datasets for model training.
- Implement low-level model changes and research patches at the PyTorch level.
- Create and verify logo and brand asset designs for internal use and customers.
- Provide customer support tasks: design edits, SVG/Lottie JSON animation edits, business card/website HTML fixes, and UX testing.
Requirements
- Proven experience training diffusion models and LLMs (both via cloud APIs and high-level libraries).
- Working knowledge and hands-on experience with Qwen-image and Z-image (Alibaba) training workflows.
- Low-level understanding of model architecture and proficiency implementing changes at the PyTorch level.
- Manual graphic/vector design skills using Adobe Illustrator or similar (non-template, non-AI tools).
- Ability to curate and prepare high-quality datasets; experience scraping and hand-curating training data.
- Strong communication and customer-facing skills to support product users while development is in progress.
Compensation & Hours
- Salary: CAD 75,000–85,000 per year.
- Full-time (listed as 30 hrs/week in posting); flexible start date.
Location & Work Arrangement
- Hybrid / Onsite in Vancouver, Canada; the position requires onsite meetings and in-person work.
Notes
- Small team environment; the role requires versatility and willingness to work across ML engineering, design, and support tasks.