ML Engineer, Trust & Safety [Hapiko].

Accountabilities:

Own and maintain the end-to-end content moderation pipeline, ensuring safety at every stage, from children's voice requests to the final AI-generated images and printed stickers.
Develop and maintain technical safeguards for AI-generated content, including the detection and handling of potentially harmful, inappropriate, copyrighted, or licensed character-related material.
Build, expand, and maintain safety evaluation frameworks to measure moderation performance and assess the impact of changes to models, prompts, and classification thresholds.
Define and manage evaluation criteria to determine whether new models, prompts, and moderation configurations meet safety standards before being deployed to production.
Continuously improve machine learning classifiers for harmful content detection, evaluating when internally developed models or third-party solutions are most appropriate.
Optimize moderation systems by balancing precision and recall, reducing unnecessary content blocking while maintaining strong protection for children.
Conduct ongoing red-team exercises to identify vulnerabilities, potential misuse scenarios, and weaknesses across models and moderation workflows.
Develop and maintain internal review tools that enable teams to investigate flagged content, assess moderation decisions, and respond to parent reports.
Own and continuously improve incident response procedures and operational runbooks for safety-related issues.
Manage the technical implementation of compliance requirements, including data retention, deletion, and other relevant data protection practices.
Collaborate closely with Machine Learning, Backend, and Platform Engineering teams to integrate safety requirements into new models, product features, and technical infrastructure before release.
Monitor the effectiveness of existing safeguards, identify emerging risks, and implement continuous improvements as the product expands internationally.
Establish and maintain reliable, scalable, and fail-safe moderation systems that support a growing user base and evolving AI capabilities.
Requirements:
Demonstrated professional experience in applied Trust & Safety engineering, particularly developing safety solutions for products used by real customers.
Strong software engineering expertise in Python, with the ability to develop, deploy, and maintain production-grade code within performance-sensitive moderation pipelines.
Proven experience building and improving machine learning classifiers or evaluation frameworks for harmful content detection, rather than simply configuring third-party moderation APIs.
Solid understanding of content moderation methodologies, machine learning safety systems, and techniques for identifying inappropriate or harmful AI-generated content.
Practical experience designing and maintaining evaluation suites to measure model performance, validate safety improvements, and assess the impact of technical changes.
Strong understanding of precision and recall trade-offs, with the ability to establish appropriate classification thresholds and balance false positives against false negatives.
Experience optimizing moderation systems to minimize unnecessary restrictions while maintaining effective protection against harmful content.
Familiarity with red-teaming methodologies, adversarial testing, and identifying vulnerabilities in AI models and content generation pipelines.
Ability to develop reliable, efficient, and fail-safe production systems, particularly in environments where incorrect moderation decisions can have significant consequences.
Understanding of incident response practices, safety investigations, and operational procedures for managing sensitive content-related issues.
Familiarity with data privacy, retention, deletion, and technical compliance requirements.
Strong collaboration skills, with the ability to work effectively alongside machine learning, backend, and platform engineering teams.
Excellent analytical and problem-solving abilities, combined with sound technical judgment and the capacity to make informed decisions under pressure.
A proactive and ownership-driven mindset, with the ability to independently lead critical engineering initiatives and establish technical processes from the ground up.
Genuine interest in responsible AI development and creating safe, trustworthy digital experiences for children.
Benefits:
Competitive compensation: Annual salary ranging from CA$136,700 to CA$182,300 for Ontario-based candidates, depending on experience, qualifications, and relevant expertise.
Remote working opportunities: Remote-first arrangements, with the advertised opportunity available in Toronto.
Flexible working hours: Flexibility to organize your schedule and support a healthy work-life balance.
Comprehensive benefits: Access to a broad benefits package.
Career development: Opportunities for professional growth, career progression, and internal mobility.
Innovation-driven environment: Work on pioneering AI-powered technologies and contribute to the development of safe digital experiences for children.
Collaborative culture: Join a hands-on, dynamic engineering environment that encourages creativity, teamwork, and meaningful contributions.
Professional impact: Take ownership of a critical Trust & Safety engineering function and influence the development of safety standards and infrastructure.
Inclusive workplace: Join an organization committed to accessibility, diversity, and equal employment opportunities.
Engaging work environment: Be part of a creative and fast-paced team that values innovation, collaboration, and fun.