Machine Learning Engineer - Trust & Safety at Hapiko (Spin Master)

About Hapiko and Spin Master

Hapiko, a Spin Master company, develops Stickerbox, an innovative voice-activated sticker printer designed for children. This product allows kids to voice their imagination and receive real printed stickers instantly. Backed by Spin Master, a global leader in children’s toys, entertainment, and digital gaming, this small senior engineering team aims to pioneer the future of play.

Role Overview

As the first Trust & Safety Engineer on the team, you will be responsible for guaranteeing that every sticker a child prints is safe. Given that the product leverages generative AI in children's environments, safety is embedded as a fundamental element rather than an afterthought. You will take full ownership of the moderation pipeline, evaluation frameworks, and related tools, ensuring robust protection while allowing kids' harmless creativity to flourish.

Key Responsibilities

• Manage and enhance the end-to-end content moderation system covering all stages from children’s voice inputs to sticker image outputs, including handling copyrighted and licensed content.
• Develop and maintain safety evaluation suites to assess whether updates in models, prompts, or thresholds meet safety standards before deployment.
• Continuously improve classification systems by determining appropriate use of in-house and third-party models.
• Fine-tune moderation parameters to reduce unnecessary blocking without compromising safety.
• Conduct ongoing red teaming exercises against models and moderation pipelines to identify vulnerabilities.
• Maintain the tooling for reviewing flagged content and managing parent reports.
• Develop and update incident response procedures for safety-related issues.
• Ensure compliance-related technical aspects such as data retention and deletion policies are upheld.
• Collaborate closely with Machine Learning and Backend/Platform teams to integrate safety measures into new models and features prior to release.

Candidate Profile and Skills

• Experience applying safety engineering principles within products used by real users.
• Hands-on experience building classification or evaluation systems for harmful content rather than solely relying on vendor APIs.
• Proficiency in Python programming with ability to deliver production-grade, efficient, and fail-safe code in moderation systems.
• Understanding of precision and recall trade-offs, and capacity to make informed decisions setting thresholds balancing error costs.
• Ability to exercise sound judgment under pressure.

Compensation and Benefits

For Ontario-based candidates, annual compensation ranges from $136,700 to $182,300, determined by factors including relevant experience and education. Spin Master offers a dynamic work environment with career growth, flexible hours, collaborative culture, comprehensive benefits, and various additional perks.

Additional Information

Spin Master encourages applications from all backgrounds and is committed to accessible and inclusive hiring, providing reasonable accommodations upon request. This position is based in Toronto, Ontario, Canada and requires onsite presence. Internal applicants should apply through the company's internal system.