Multi-modal Machine Learning Engineer (all levels) - Trust and Safety

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

Builds machine learning models and serving systems that moderate harmful content across a global short-video platform at very large scale. The Trust and Safety team in this role is fast growing and aims to protect users and publishers from negative content, with candidates evaluated across five sub-teams.

What you'd do

• Lead and build content understanding models used for content moderation.
• Build classifiers over text, images, and video, plus graph, representation learning, and LLM-based models.
• Deploy machine learning models directly to the global platform and design online model workflows.
• Implement serving pipelines that run at scale across the platform.
• Work with product teams to define business objectives and shape trust and safety strategy.
• Team up with data scientists to pin down hard problems and suggest data-led approaches.

What it asks for

• BS or MS in computer science, statistics, operations research, applied mathematics, physics, or a similar quantitative field.
• Hands-on experience in one or more of: CV, NLP, multimodal, graph, or LLM machine learning, statistical models, or applied mathematical methods.
• Strong coding skills, especially in Python or similar languages.
• Several years of building production software beyond the modeling work.
• Ph.D. in LLM, VLM, or equivalent experience (preferred).
• Background in LLM, multimodal modeling, or representation learning, or publications at AI/ML conferences (preferred).
• Grasp of how content understanding and moderation create business value (preferred).

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.