Applied Scientist Intern - Trust and Safety (Multimodal Foundation Model) - Global Frontier Tech Recruitment Program - 2027 Start (PhD)

TikTok
Seattle
Workplace: OnsiteInternshipFunction: Research & Scientific (R&D)Education: phdSkills: ["Problem-solving","Creative mindset","Research execution","Critical thinking"]

Build frontier multimodal trust & safety models and systems to protect users from negative content. Work on multimodal moderation foundation models (large-scale MoE training, cross-modal alignment, synthetic data generation) and agentic moderation systems using reinforcement learning for multi-step decision-making, tool ecosystems, and interpretable closed-loop reasoning. Tackle challenges like training stability, routing optimization, context engineering, GraphRAG retrieval, and adversarial robustness across 200+ languages.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
TikTok
TikTok
1 month ago

Applied Scientist Intern - Trust and Safety (Multimodal Foundation Model) - Global Frontier Tech Recruitment Program - 2027 Start (PhD)

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 23 hours agoStatus: Live

Job Summary

Build frontier multimodal trust & safety models and systems to protect users from negative content. Work on multimodal moderation foundation models (large-scale MoE training, cross-modal alignment, synthetic data generation) and agentic moderation systems using reinforcement learning for multi-step decision-making, tool ecosystems, and interpretable closed-loop reasoning. Tackle challenges like training stability, routing optimization, context engineering, GraphRAG retrieval, and adversarial robustness across 200+ languages.
Location: Seattle
Workplace: Onsite
Employment Type: Internship
Job Function: Research & Scientific (R&D)
Seniority: Intern level

Key Responsibilities

  • •Develop multimodal moderation foundation model approaches, including large-scale MoE architecture training and routing optimization with cross-modal alignment and reasoning.
  • •Build and improve agentic moderation systems using reinforcement learning for multi-step decision-making and tool-call strategies.
  • •Design context engineering and tool collaboration components, including dynamic context assembly and MCP-based tool ecosystem construction.
  • •Implement evaluation and robustness improvements, including generalization across many languages/strategies and adversarial detection for AIGC content.
  • •Advance research from idea to implementation, producing tangible outcomes in scalable content safety systems.

Key Requirements

  • •Currently pursuing a PhD in Computer Science, Data Science, Artificial Intelligence, or a related field.
  • •Proficiency in programming languages such as Python, Rust, or C++ and experience with deep learning frameworks (e.g., PyTorch, DeepSpeed, Megatron, vLLM).
  • •Strong understanding of distributed computing, performance tuning, and verification for training, finetuning, and inference.
  • •Experience with inference tuning and inference acceleration, including deep understanding of GPU and/or other AI accelerators.
  • •Strong understanding of cutting-edge LLM research including long context, multi-modality, and alignment, with practical expertise implementing advanced systems.
Education:PhD / Doctorate
Skills:Problem-solvingCreative mindsetResearch executionCritical thinking
Tech Stack:PythonRustC++PyTorchDeepSpeedMegatronVLLMGRPOPPOMCPGraphRAGGPUPEFTRLMoECoTLangChain

Company Brief

TikTok
Short-form video platform that lets users create, share, and discover entertainment content through algorithmic recommendations. It also offers advertising and creator tools for brands, influencers, and businesses.
Industry: Digital Media
Company Size: Enterprise (1,001+ employees)
Growth: Scaleup
Headquarters: Singapore, Singapore
Founded: 2016
WebsiteLinkedIn