Diffusion ML Engineer: Character Models for Kids

Spin Master; Toronto; Hybrid; CAD 137,000 - 182,000; Full time

Benefits offered by this job

Growth opportunities

Flexible hours

Comprehensive benefits

Innovation & collaboration

Fun perks

Job summary

Spin Master Hapiko is seeking a hands-on model engineer to build and tune diffusion models for a first-of-its-kind voice-activated sticker printer for kids. You will train data, evaluate outputs, and work with brand teams to ensure on-model accuracy and safety.

You’ll join a small team inside Spin Master, collaborating with product and safety, shipping updates globally while keeping kid-friendly outputs and compliance at the forefront.

Qualifications

• Hands-on experience fine-tuning diffusion or generative image models.
• Built training data pipelines with provenance requirements.
• Built evals for generative output to inform decisions.
• Strong Python software engineering for reproducible pipelines.

Responsibilities

• Train and fine-tune diffusion models, pipelines, and adapters for characters.
• Extend evaluation harness to measure character fidelity alongside output quality.
• Ensure a kid’s spoken request yields the right outputs.
• Work with brand teams to get characters approved and shipped.
• Partner with ML Serving and Trust & Safety engineers to productionize products for kids.

Skills

Diffusion models

Training data pipelines

Model evaluation

Python engineering

Curiosity

Job description

Spin Master Hapiko is seeking a hands-on model engineer to build and tune diffusion models for a first-of-its-kind voice-activated sticker printer for kids. You will train data, evaluate outputs, and work with brand teams to ensure on-model accuracy and safety.

You’ll join a small team inside Spin Master, collaborating with product and safety, shipping updates globally while keeping kid-friendly outputs and compliance at the forefront.