Machine Learning Engineer

About Vectica Systems - Brandmark.io

We are a (very) small team based in Vancouver, Canada creating the next generation of AI-powered logo design tools and products. We aim to live up to our namesake and deliver a full AI-assisted brand identity design experience beyond just a logo.

We believe that AI has the power to empower everyday users to fulfill their brand vision at an affordable cost, without expensive subscriptions.

About the Role

We are seeking a versatile Machine Learning Engineer to help develop the next generation of our brand identity design chatbot and AI image generation products. The position will require onsite meetings in our Vancouver location.

Full-time · 30 hrs/week | Hybrid / Onsite — Vancouver | $75,000–$85,000 CAD/year | Flexible start date

Because we are a (very) small team the role will require wearing multiple hats, including:

• Training of LLMs and diffusion models
• Scraping/hand curating data for training of said LLMs and diffusion models
• Logo design and brand asset design for both internal use and for brandmark.io customers
• Customer support for brandmark.io customers

Skills & Experience

[Required] Logo / Graphic Design Training diffusion models for logo design requires proficiency in logo/graphic design, not only to verify the quality of the resulting model but throughout the training process from data selection to model architecture. The taste and aesthetic preferences of the human are reflected in the final model. The ideal candidate will have *manual graphic design skills* using Adobe Illustrator or similar vector design tool (non-Canva, non-template, non-AI design tools).

[Required] Familiarity with Qwen / Z-image Although there are other open-weight models (Stable Diffusion / Flux etc.), at brandmark.io we use Qwen-image and Z-image from Alibaba. These are the best unencumbered open source models at the moment and we expect to continue to use them in the future. Working knowledge of Qwen / Z-image training is *required*, both for LoRA training as well as full-weight training. The ideal candidate will have experience not only with training Qwen-image using high level libraries but the model architecture at the PyTorch level. (No vibe coding.)

[Required] Training of LLMs Future software interfaces will use chat and agentic workflows as a matter of course. We will need to train a domain-specific LLM specifically for brand identity design. Required: Experience training LLMs both via cloud APIs and via high-level libraries like Hugging Face Transformers. The ideal candidate will have low-level understanding of LLM architecture at the PyTorch level and have the (manual) coding skills required to implement the latest transformer research by patching the training/inference engine code.

Customer Service & Communication While models are training and code is compiling you may be tasked with serving brandmark.io client requests. This may involve designing customer logos/graphical assets, updating or fixing existing customer logos/graphics, editing SVG/Lottie JSON animations, editing customer business card/website HTML/presentation asset downloads, and debugging/UX testing of the brandmark.io website on desktop and mobile platforms.

Pay: $75,000.00-$85,000.00 per year

Work Location: In person