ML Engineer, Trust & Safety [Hapiko]

Role overview Build and operate the technical safeguards for a generative AI product used by children. As the first dedicated trust and safety engineer, you will own the moderation pipeline and its evaluation and review tools, balancing strong protection with allowing harmless creative ideas through. The work is hands-on software and model evaluation, with safety integrated into new features from the start.

Responsibilities - Maintain moderation across the full flow, from a child’s spoken request to the image that is ultimately printed, including handling copyrighted or licensed characters. - Build and expand safety evaluation suites to assess changes to models, prompts, and moderation thresholds before release. - Improve classifiers and determine when in-house or third-party models are appropriate. - Tune moderation to reduce unnecessary blocking while maintaining effective safeguards. - Conduct ongoing red-team exercises and maintain tools for reviewing flagged content and parent reports. - Own the technical incident response process and support compliance needs such as data retention and deletion. - Partner with machine learning and backend engineers to incorporate safety into models and features before launch.

Requirements - Applied trust and safety engineering experience on a product with real users. - Experience building content safety evaluations or classifiers, beyond configuring a vendor service. - Strong Python engineering skills and ability to ship fast, fail-safe production code. - Sound judgment balancing precision and recall when both missed harmful content and over-blocking have consequences. - Ability to make decisions under pressure.

Benefits and work setup The role is listed as remote. The package includes flexible work hours, comprehensive benefits, and career development opportunities.