Technical Product Manager, GenAI Safety Solutions

Written by the T&S Network from TikTok's posting.

Builds the shared GenAI safety platform and strategy for a Trust & Safety product team, covering evaluation, moderation and guardrails across GenAI models and applications. The team exists to ensure GenAI launches are safe, tested and managed for risk before release.

• AI misuse and jailbreaks

What you'd do

• Build features for centralized GenAI safety platforms, including evaluation and auto-labeling, a unified moderation pipeline, and post-launch monitoring.
• Develop GenAI content moderation models with Algo teams for content safety and IP/copyright risk across modalities.
• Make safety filters and cleaning pipelines for foundation models.
• Research emerging GenAI risks such as jailbreaks and attack methods, and turn findings into platform safety features.
• Oversee long-term safety solution programs with stakeholders and cross-functional teams.

What it asks for

• Five or more years of work in product, engineering, data science or machine learning.
• Proven experience building solutions from the ground up, alone or in a small team across product, R&D, Algo and DS.
• Fluent Mandarin and fluent English.
• Product design skills and comfort working with data on complex products.
• Hands-on experience with GenAI and AI-assisted agentic development tools.
• Trust & Safety background (preferred)
• Go-getter mindset for overcoming business challenges (preferred)

Terms in this role

Red teaming.

Testing a system, often an AI model, by deliberately trying to make it fail or misbehave, so problems are found and fixed first.

Jailbreak.

A prompt written to get an AI model to ignore its safety rules.

Our best summary of the posting, and it can be wrong. The employer's posting is what counts: read it in full on TikToknorth_east before you apply. If we got something wrong, report this listing.